Bf 109 G-2 climb data

[Holtzauge]

The diagram shows C++ simulation climb rates for different 109’s as a function of power to weight ratio (P/W). As can be seen, there is a good correlation between P/W ratio and climb speed which also seems reasonable given that the aircraft are so similar.

 

Granted, these are calculated values but with the exception of the G2 I think the correlation with historical data is satisfactory. For some reason the available G1 and G2 data on climb rate at 2 Km altitude has a great spread for the 1.3 ata power setting: On the high end we have the Finnish test at 24 m/s and Erla 21 m/s and on the lower end figures around 17 m/s all of which can be found at WWIIaircraftperformance.org.

 

Given that there is more data available and more concensus around the F4, G6 and K4 figures, I’m inclined to trust the simulation figure of 18.7 m/s for the G2 as more representative. This simply makes sense given the G2’s placement in terms of P/W ratio compared to the other 109’s and also fits somewhere in the middle of the available data.

 

Concerning the Finns climb test of the MT-215, they say in the report that the speed test is in their opinion accurate but that the climb tests is ”somewhat innacurate” and that no special flight instumentation was available for the test. In addition, seeing that the trial lasted more than 45 minutes it would also be interesting to know MT-215’s fuel status when they did the climb test since this will have a significant impact on the results and could compound any measurement error.

 

As an added benefit of doing the G2 modelling, I can add that the calculated acceleration for the G2 from 300 to 440 Km/h is pretty much the same as for the F4 with the G2 needing 26.3 s as compared to the F4 27.0 s.

 

BTW: Has someone tested the ingame G2 climb rate at 2 Km altitude?

Me109climbcomp.bmp

[Kurfurst]

Comparing this estimation of climb performance to real life tests, all in considerably different airframe (+/- 50 km/h worth of drag - "so similar" :D ) and powerplant condition is a basically pointless excercise IMO. There is just too much variation in the condition of the base IRL data.

 

p.s. 1,98 ata K-4 on Mike's site? :D

Edited May 28, 2014 by VO101Kurfurst

[Holtzauge]

Well hopefully the developers will not be discouraged by a variance in the data because in the end the modelling will have to be made on the basis of some IRL data. The problem right now seems to be that the IRL Bf109G2 data spans from about 17 m/s to 24 m/s. The question is which is representative?

 

In the paralell thread on G2 acceleration, the following ingame performance has been measured:

 

300 to 400 km/h acceleration time : 17.6 s

440 to 510 Km/h acceleration time : 18.0 s

 

Now the performance I get in the C++ simulation (at 15 deg C) for the Bf109G2 is the following:

 

300 to 400 km/h acceleration time : 26.3 s

440 to 510 Km/h acceleration time : 31.8 s

 

This is pretty close to the Finns MT-215 measurements which seem to have been done at 11 deg C. In addition, while the temperature does have an impact, the delta in power between 11 and 15 deg C is most likely in the order of 20 Ps (see reasoning in acceleration thread) so the conditions between the simulation and Finnish tests should be comparable.

 

Now in order to get the ingame acceleration times listed above, the engine power in the C++ simulation has to be souped up and when the ingame acceleration times match the times I get in the simulation I get a climb rate of 24.5 m/s at 2 Km.

 

So based on this it looks like the developers have used the MT-215 climb rate of 24.5 m/s as input to the G2 modelling which resulted in the very high ingame acceleration times

 

But using the MT-215 acceleration times instead, the C++ simulation gives a climb rate of 18.7 m/s at 2 Km.

 

So my input to the delvelopers is that 18 to 19 m/s not 24.5 m/s in standard atmospheric conditions would be a more reasonable climb rate at 2 Km.

Edited May 31, 2014 by Holtzauge

[Kurfurst]

Well hopefully the developers will not be discouraged by a variance in the data because in the end the modelling will have to be made on the basis of some IRL data. The problem right now seems to be that the IRL Bf109G2 data spans from about 17 m/s to 24 m/s. The question is which is representative?

 

IMO the real question is: under what conditions? The IRL tests you have mentioned are practically all in different airframe etc. conditins.

 

**SNIP***

 

So my input to the delvelopers is that 18 to 19 m/s not 24.5 m/s in standard atmospheric conditions would be a more reasonable climb rate at 2 Km.

 

Again the question is: for what condition? What is the condition of the aircraft is assumed in the C++ estimation for example? Poor, equal to nominal or exceptional airframe, tailwheel retractable or not, what take off weight, power output, propellor effiency, radiator opening width and drag...?

 

Differences between results of IRL tested airframes can be all explained by these conditons. If your C++ estimation however ignores these and assumes that all aiframes were in similar condition, do not be surprised that the results of IRL tested airframes that were actually in very different shape and testing condition than your model yielded different results. 

 

MT 215 climb results are, for example, perfeclty normal and fit very well to other known tests, if the testing prototocol deviation in that tests (closed radiators at the beginning of the climb) are taken into account. The radiator position of the 109 had a strong influence on drag and rate of climb.

[Holtzauge]

No, the MT-215 climb data cannot be explained by any of the things you listed: Begining with the conditions, the weight condition used in all the C++ comparisons is the same: All aircraft are fully armed and fueled. But that is really not an issue because the climb rate is determined by the thrust to weight ratio and the first figure I posted as an attachement here can be used to gauge how increased power or reduced weight affects the G2.

 

I did some complementary C++ simulations just to illustrate the point: Beginning with weight, assuming the the G2 is running on fumes at a weight of 3030 Kg minus 320 Kg fuel weight = 2710 Kg this gives a climb rate of 21.5 m/s. This gives a P/W ratio of around 0.52. So even with an empty tank it does not reach 24.5 m/s climb rate. If however, we assume a G2 on steroids with a 25% power increase then we reach 24.5 m/s. This gives a a P/W ratio of around 0.58.

 

Now if you and Crump look in the figure I posted and read off the climb rate at those P/W ratios you will see that they correspond pretty well to the climb rates of 21.5 and 24.5 m/s, i.e. the figure and the Me109F4 can actually be used as a basis to gauge the climb rate of the G2.

 

Now concerning the radiator drag. I have assumed the best condition possible so the radiator is fully closed for all aircraft in the comparison. With a fully open radiator, it is top speed that suffers most but climb rate is of course also reduced but only by about 1m/s. So in that case the original 18-19 m/s estimate is reduced to 17-18 m/s. The effects of airframe condition and a retractable tail wheel on climb are marginal.

 

BOS obviously has a flight model and so do I. The problem here is that with a proper flight model it is not possible to reconcile the 24.5 m/s climb data with the G2 P/W ratio. Again, my post is directed at the developers and frankly I will leave it at that because I think they understand what I'm getting at.

 

From experience I know you have a penchant for advocating outliers in the data on the 109 and doggedly defending them and If you want to continue to try to convince them that "MT 215 climb tests are, for example, perfectly normal and fit very well to other known tests" then I wish you luck with that and look forward to seeing that data but I don't think they will be swayed towards a 24.5 m/s climb rate by arguments put forward so far.

[Cpt_Branko]

Question, what methodology do you use to obtain thrust?

Edited June 5, 2014 by Cpt_Branko

[Cpt_Branko]

I would be wary of using tests of captured machines (many can be found on Williams' site), and use a more accurate indication of climb: time to altitude.

 

With this, we see that there is in fact not much deviation, eg:

                     Time to climb

1km             0.8 min (MTT-215)      0:78 (Reichlin)

2km             n/a                             

3km             2.3 min (MTT-215)      2:4 (Reichlin)

4km             3.2 min (MTT-215)      3:27 (Reichlin)

5km             4.1 min (MTT-215)      4.18 (Reichlin)   4.4 (NI VVS)  4.4 (Erla)

 

When we focus on time to climb test, the 4:18 vs 4.1 minutes is a nearly insignificant difference  (less then 2%) and easily within limits of experimental error, easily attributed to slight weight differences (Reichlin and Erla tests were on G-1 machine, Erla machine seems to have GM-1 booster which is not useful at altitude below 5km but does increase weight a bit) and radiator position differences. Agreement within 2% is rarely there for WW2 performance trials I've seen.

 

In fact taking the NI VVS test - and it has to be said, this is a captured machine so the state of the airframe is not ideal (although it is the most decently performing captured 109 test I have seen), the time to climb to 5km in NII VVS test is 4.4 minutes (a difference of 7%, which is really good for a test of a captured machine). In ERLA test, I'm again reading out 4.4 minutes to 5km from the graph, again a 7% difference - but bear in mind, again, that it is a G-1 model with GM-1 booster also, slightly increased weight is to be expected.

 

So basically, all the data coming from Germany, including data from Finland but also from VVS is in good agreement with each other. It might not look that way when we look at maximum m/s obtained, but that is almost a useless metric and a small error in measurement could lead to a significant error in numbers at some narrow altitude band (which is why, time to climb is a much more "reliable" figure for making comparisons).

 

However, it is good that people run calculations and try to model performance

Edited June 6, 2014 by Cpt_Branko

[Kurfurst]

Details and translation of the Finnish Air Force trials with the imported Bf 109G-2 "MT-215" can be read here.

 

http://kurfurst.org/Performance_tests/109G_MT215/109G2_MT215_en.html

 

There also a detailed explanation of the reasons behind the exceptional climb performance of the aircraft between ca 0-2 km altitude. In short, the report notes that the climb trials were conducted with greater than prescribed airspeed and that the radiators were closed up to ca 2500 m altitude (the result of low coolant temperatures and thermostatically controlled radiator kinematics). The closed radiators considerably reduced the drag of the aircraft (this is confirmed by German trials with the 109 BTW), which is why the climb is greater up to the altitude of 2500 m, when the radiators open and the climb rate is reduced. Above that altitude the climb results are in very close agreement with German and Soviet trials of similar model aircraft, as Branko noted.

 

From my previous testing and consultancy experience, one of the major modelling problems with the original Il-2 was the incorrect drag modelleing of radiator drag on the 109. I believe this is the issue with the C++ calculated model as well.

[Cpt_Branko]

Even time to 2km (well, Finnish test is timed at 1.9 km) is not dramatically different to, say, Reichlin trials. 1.58 to 2km versus 1.4 to 1.9km is not a dramatic difference (corrected for altitude, about 7.5%) considering the radiator position differences.

 

In general, if we talk about working out climb rates, it's important to note, increases in P/W are going to not going to linearly translate to increases in climbrate. Rate of climb is directly proportional to

 

(Power_available - Power_required) / Weight

 

Available power is basically engine power * propeller efficiency (while engine power is generally a "flat" line regardess of speed (simplification, but valid enough), power available is not a flat line for propeller driven aircraft, because propeller efficiency varies with speed).

 

A direct increase in engine power (assuming propeller efficiency stays the same at maximum power setting, which is really a big assumption) would produce a nonlinear difference (although; the power required part might see some minor changes, as well, due to an increase in propeller drag). The same way - reduction or increase in weight will also not produce an exactly linear difference in rate of climb, because with weight, drag also changes (total drag is a consequence of both induced and parasitic and induced is very much a function of L, which is a function of W in steady flight).

 

Take a look at Aerodynamics for Naval aviatiors, around pages 150-155.

Edited June 6, 2014 by Cpt_Branko

[Holtzauge]

Cpt_Branko: I agree that using data from captured aircraft can be misleading. However, I see that you refer to data on WW2Aircraftperformance site and actually there is data from Messerschmitt documents there as well that indicate climb rates substantially lower than the ones you mention.

 

I also agree that point data on climb rate may be misleading so going with climb times instead of maximum climb rates:

 

There is one (Erla) giving a climb time of 1 min 35 s to 2000 m, i.e. circa 21 m/s

 

Another (Messerschmitt document) which gives 2 min to 2000 m, i.e. circa 16.7 m/s

 

Yet another Messerschmitt document gives 1.75 min to 2000 m, i.e. circa 19 m/s

 

So there is a great spread in the data. In addition, note that (as I posted above) the Finns themselves noted that they considered the MT-215 speed tests reliable while the climb test was termed ”somewhat innacurate”.

 

In the climb data table you posted Cpt_Branko, I noticed that the best Rechlin climb rate (apart from the Finnish MT-215 data) is in the order of 21m/s which is a huge improvement from assuming a climb rate of 24.6 m/s and getting closer to realistic values IMHO.

 

As I explained above, the good thing with a good flight model is that if you have reliable data for some points in the performance envelope (e.g. speed, climb, turn rate or accelereration) then you can make good predictions for the others. The problem here is that if I tune my C++ model for the high climb rate then the other performance parameters go off the chart:

 

My best estimate tuned Me109G2 model gives the following performance at 1.3 ata and 3030 Kg weight:

 

522 km/h at 0 m

 

646 Km/h at 6.7 Km

 

Accelereration:

 

300-440 Km/h: 26.3 s

 

400-510 m/h: 31.76 s

 

Best turnrate at 1 Km alt: 20.16 s at 280 Km/h speed

 

Climb rate at 2 Km/h: 18.69 m/s

 

So comparing this to the Finnish MT-215 speed tests its close. Also comparing to the highest German speed estimates its close. The acceleration figures are also close to the Finnish tests, especially seeing I don’t model engine reving up so being picky one should add one or two s to my figures which brings them even closer. The turnrate is also close to the figures I’ve seen. In addition, I get almost exactly the same turn time as the Finns at 360 Km/h, namely 22.7 s.

 

The only thing here that really deviates is the climb rate which the Finns themselves labeled unreliable.

 

Going the other way and tuning the C++ model for the 24.6 m/s climb rate at 2000 m:

 

563 km/h at 0 m

 

699 Km/h at 6.9 Km

 

Accelereration:

 

300-440 Km/h: 18.69 s

 

400-510 Km/h: 17.08 s

 

Best turnrate at 1 Km alt: 18.14 s at 300 Km/h speed

 

Climb rate at 2 Km/h: 24.6 m/s

 

Now which makes more sense? These speed and climb figures are more reminicent of the Me109K4 at 1.8 ata than they are of the G2 at 1.3 ata.

 

The point I have been trying to make is that since you see the acceleration times you do in BOS, it is most likely so that the G2 model is overly optimistic. However, until the developers shed some light on this or someone actually measures the ingame climb speed we don’t know but based on the acceleration times it certainly seems so.

 

@Kurfurst: Do you have some data on how the climb rate is affected by the radiator flap position to the extent you think? There is some data on WW2aircraftperformance on a G6 which indicates around 1 m/s difference between fully closed and fully open at 2 Km altitude. Do you have other data that indicates a higher difference?

 

@Cpt_Branko: Concerning the propeller efficiency, I have this varying with speed, prop blade loading, tip mach and the engine power and critical altitude dependant on ram pressure etc and spent some time tuning so I think its pretty close. In addition, the climb rate method you refer to only using power difference (SEP climb rate) is valid only for small climb angles. WW2 figthers climb at a steeper angle and this should be taken into account. Also, I see you are saying that the chart I posted which shows a close to linear relationship between P/W ratio is not valid. I’m not saying that there necessarily is perfect linear relationship. I’m showing the results for different aircraft and pointing out that it seems close enough to use for comparisons. If the figure is off what should it look like? Which Me109 versions in the chart are misrepresented?

[Cpt_Branko]

Interesting that all the other numbers match relatively well (with some differences depending what document you take, for instance - Soviet test of G2 quotes turn time to be 20s / 21.5s depending on side (left / right turns) at 1km altitude).

 

Sadly I don't have any real data on VDM propellers or detailed engine performance, only have it for Allied types (if you ever want to try calculating for those, I can share what documentation I have on them), so I can't help there.

 

What would the time to climb in your simulation be up to, say, 2, 3, 5km? I think it gives a best comparison of overall climb rate, comparing the "peak values" might give a skewered idea of climb rate. I'm interested if the overall times match more then "peak" rate which might involve an error in measurement of the day (but doing it to 5km, they tend to get averaged out, and most tests are in very good agreement in time to climb to, say, 5km - while they vary a bit more in peak climb rates in specific altitude bands).

 

As for the graphs not really following P/W (when data from German sources is used), this is what I mean:

Bf-109E, from manual, avrg. climbrate between 1-2km: 18.51m/s, avrg. P/W between 1-2km (from my understanding of the chart, 1000 HP as tested, but without engine chart possible variation) & 2540kg: 0.394 hp/kg

 

Bf-109F4: from datenblatt derived from flight tested performance, avrg. climbrate between 1-2km: 18.51m/s, avrg. P/W between 1-2km on steig-u-kampfsteilung (1250 HP, but without engine chart possible variation) & 2890kg: 0.433 hp/kg

 

Bf-109G1, from datenblatt (derived from calculation), avrg. climbrate between 1-2km: 18.51m/s, avrg. P/W between 1-2km on steig-u-kampfsteilung (1310 HP, but without engine chart possible variation) & 3042kg: 0.431 hp/kg

 

Reichlin figures give 20.83m/s from 1-2km for the same Bf-109G1 (10 seconds less), Erla chart for G1 reads out at about 19.5m/s (flying weight ~3070kg), Soviets quote 19m/s for G2 at those altitudes.

 

When I look at MT-215 results, yeah, they do stand out a lot in that specific region between 1-2km, but overall climb time to say 5km is similar within limits of experimental error with Reichlin figures, and very close to Soviet figures. It might be a case of not so reliable climb measurement in specific climb regimes.

Edited June 6, 2014 by Cpt_Branko

[Holtzauge]

OK, thanks for the input. I don't have time right now but I will post climb data with altitude later today ;-)

[Holtzauge]

Attached is a a chart showing the C++ simulated climb speed as a function of altitude for the standard Me109G2 I listed performance for above. IMHO it seems more reasonable than the high end German and Finnish MT-215 data given that this is what you get when you tune in the other parameters like, top speed, acceleration and turn rate.

Looking at the MT-215 climb chart, it seems extrapolated from a few data points that can be seen in the climb chart and which the Finns themselves termed as only "somewhat reliable". I feel a need to reiterate this point because the chart keeps being quoted while the Finns themselves were specific that they considered only the speed test (which coincides with my data) reliable and the climb tests unreliable since no special instrumentation had been used.

I also did an integration of the climb speeds and got the following results:

Climb time to 2 Km altitude: 1.83 minutes i.e. around 18.18 m/s

Climb time to 5 Km altitude: 4.98 minutes i.e. around 16.77 m/s

Cpt_Branko, I noticed you focused on the high end Finnish and Erla/Rechlin data, but there is also other Messerschmitt data (i.e. not from captured aircraft) that gives lower and a couple of charts give the climb rate to 5 Km altitude at 5.2 minutes i.e. around 16 m/s in average. I think the Erla/Rechlin climb times of 21 m/s at low altitude are optimistic. In fact the whole Erla/Rechlin table seems overly optimistic since they list a climb speed of 1 m/s at 12000 m altitude. Seeing the ceiling (defined by 0.5 m/s climb remaining) of the G2 is more likely in the neighbourhood of 11500 m I think this whole table seems on the optimistic side. On the other hand, the data you posted on the climb rates at 1-2 Km altitude are close to my estimates so I think we have good agreement there and I see your point about the P/W varying with altitude but the chart I posted in the first post is only valid for 2 Km altitude and was done to illustrate the 24.6 m/s climb anomaly.
 

Anyway Cpt_Branko, I'm happy to see that we agree that 24.6 m/s at 2 Km altitude is overly optimistic. The question now is to determine what is reasonable?  

Me109G2 climb altitude chart 1.bmp

Edited June 6, 2014 by Holtzauge

[Holtzauge]

Correction: Made a mistake    in the integration of the C++ climb rates above. Just integrated that code into the model and I did not compensate Vt with altitude to mainatin a suitable Vi so climb rates in the integration were non-optimal. New estimates are:

 

Climb time to 5 Km 4.79 minutes i.e. an average climb rate of 17.42 m/s

 

Climb time to 2 Km 109 seconds i.e. an average climb rate of 18.35 m/s

[Crump]

Comparing this estimation of climb performance to real life tests, all in considerably different airframe (+/- 50 km/h worth of drag - "so similar" :D ) and powerplant condition is a basically pointless excercise IMO. There is just too much variation in the condition of the base IRL data.

 

p.s. 1,98 ata K-4 on Mike's site? :D

 

The difference between recorded data from flight testing and this "simulation" is probably due to the underestimation of propeller efficiency in the "simulation".

 

Cpt_Branko says:

 

Interesting that all the other numbers match relatively well (with some differences depending what document you take, for instance - Soviet test of G2 quotes turn time to be 20s / 21.5s depending on side (left / right turns) at 1km altitude).

 

Sadly I don't have any real data on VDM propellers or detailed engine performance, only have it for Allied types (if you ever want to try calculating for those, I can share what documentation I have on them), so I can't help there.

 

You are on the right track!

 

There is not much to be said.

[novicebutdeadly]

Gents,

 

Since I've got 5 cents in my pocket (we don't have 2 cent coins here in Australia), I figure I might throw my opinion in.

 

Off the top of my head:

I thought the G2 in the soviet test had under wing guns installed??????????
 

From memory the allies tested a captured G2 in North Africa, the aircraft had a damaged/defective radiator control which resulted in the radiators being stuck open.

The noted climb rate of this aircraft was 20m/s

So if a G2 with a trop filter, and with radiators stuck open can climb at 20m/s, I don't think it's out of the ordinary for an initial climb rate of circa 22m/s.


However things that come into play (not already mentioned) is the performance of the engine.

In that 2 brand new power plants (same type etc) could have notable performance differences in terms of power and fuel economy depending which factory they were produced in.


Even today, there is difference in aircraft performance amongst identical aircraft, if you go to any aerodrome you will always find one plane that for reasons unknown to anyone,  will have superior performance to other identical aircraft.

This is also observable in mass produced cars, where some cars will be a little better, others a little worse.

 

The only real way to model that into a sim, is to get the average (a+b and divide by 2).

 

Edited June 9, 2014 by 907-novicebutdeadly

[Cpt_Branko]

Getting any simulation within 5-10% of expected (debatable, as warbird performance is) values is quite a feat.

 

MTT in their original perfomance specification (done based on calculation) got roughly similar values to yours (1.8 minutes to 2km, 3.6 minutes to 4km, 5.7 minutes to 6km). So your calculations aren't worse then Messerscmitt's back in the day.

 

Of course when you actually build a machine things can get better - or worse. The differences in top speed between tested and accepted machines was about 30km/h. Your calculation of top speed (646 Km/h at 6.7 Km) is within acceptable performance figures, for instance (MTT calculation 647km/h, Soviet test 650km/h, Erla average result 652km/h, but for instance machine doing 643km/h was also acceptable).

 

From flight tests, 4.4 to 5km seems to be basically the average value (Soviet test* and Erla list 4.4 - 4.5, discounting the higher Reichlin & MTT figures). If you are fluent with Russian (I am not, last time I read any cyrillic was twenty years ago), there's a document on the measurements they did on the 109G2 (their flight test results are available on Kurfurst's site, too):

https://drive.google.com/folderview?id=0B-H7PUYuIiDVYm91c2cwNkJfVWs&usp=sharing

 

 

*Soviets did tests both with and without gunpods. I'm naturally taking the ones without gunpods (taking the test with gunpods would be a bit dishonest).

[novicebutdeadly]

Getting any simulation within 5-10% of expected (debatable, as warbird performance is) values is quite a feat.

 

MTT in their original perfomance specification (done based on calculation) got roughly similar values to yours (1.8 minutes to 2km, 3.6 minutes to 4km, 5.7 minutes to 6km). So your calculations aren't worse then Messerscmitt's back in the day.

 

Of course when you actually build a machine things can get better - or worse. The differences in top speed between tested and accepted machines was about 30km/h. Your calculation of top speed (646 Km/h at 6.7 Km) is within acceptable performance figures, for instance (MTT calculation 647km/h, Soviet test 650km/h, Erla average result 652km/h, but for instance machine doing 643km/h was also acceptable).

 

From flight tests, 4.4 to 5km seems to be basically the average value (Soviet test* and Erla list 4.4 - 4.5, discounting the higher Reichlin & MTT figures). If you are fluent with Russian (I am not, last time I read any cyrillic was twenty years ago), there's a document on the measurements they did on the 109G2 (their flight test results are available on Kurfurst's site, too):

https://drive.google.com/folderview?id=0B-H7PUYuIiDVYm91c2cwNkJfVWs&usp=sharing

 

 

*Soviets did tests both with and without gunpods. I'm naturally taking the ones without gunpods (taking the test with gunpods would be a bit dishonest).

 

 

I thought they only did test with gun pods on (which did seem very bizarre to me),

 

glad to know that I'm wrong, because it means I learn something new

[Crump]

 

Cpt_Branko says:

 

Interesting that all the other numbers match relatively well

 

I would not give a second thought to this simulation and I certainly would not use it as basis to disparage the plethora of measured data that agrees.  The fault probably does not lie with the engineers at the German, Russian, or Finnish Aviation Authority.

 

Holtzauge would do better to find the issue with his assessment as to why it is pessimistic compared with the measured data points. 

Edited June 9, 2014 by Crump

[Holtzauge]

I would not give a second thought to this simulation and I certainly would not use it as basis to disparage the plethora of measured data that agrees.  The fault probably does not lie with the engineers at the German, Russian, or Finnish Aviation Authority.

 

Holtzauge would do better to find the issue with his assessment as to why it is pessimistic compared with the measured data points.

Well, I have adopted a new strategy now which is simply to ignore your silly posts but I will make an exception and make a note of your opinion on my modelling and add that to the other insights you have provided over the years like the P-47 Thunderbolt having a higher max sustained loadfactor compared to the Spitfire Mk9 at 20000 ft, that the FW190A had a Cdo lower than the Spitfire Mk9 and that the Fw190A could not fly with the outer wing guns removed due to a dangerous CG shift etc.

 

Speaking of accurate modelling, I noticed you posted a diagram on the Me109G2 turn capability in the paralell acceleration thread which shows the Me109G2 having a best turn rate of around 22.5 deg/s. I'm really impressed by this since this translates to 16 seconds to complete a full circle which seems to "disparage" historical Me109G2 data at 20 -21 seconds and even beats the Spitfire which took 17 seconds.

[Holtzauge]

Getting back on a more constructive track:

Cpt_Branko, thanks for your input. While I do agree that the climb time is an important metric, I still think the climb rate versus altitude chart is a better tool for the comparisons we are making here related to the flight model:

I added the historical climb rate data I have to the climb chart (updated attachement) and also did an integration to derive the climb time estimates. The intersting thing when comparing the historical climb times was as you pointed out that the time to altitude was close in some cases like the Rechlin and MT-215 data. However, while this could be used to say they are equal if the target is an intercept at say 6 Km altitude I think it misses the point about the difference in performance in different height bands:

The MT-215 data indicates an enormous advantage up to 3 Km altitude after which the Rechlin 109G climbes better up to 6 Km which just about evens out the climb time results. However, if we choose either model this will impact at which altitude the ingame Me109G2 will perform better: If the Finnish data is used the result will be vastly superior climb performance at 0-3 Km and poorer performance at intermediate altitudes up to 6Km, i.e. it will affect the way the aircraft performs and is used at those altitudes even if the climb rate to 5 Km altitude is about the same.

Again, IMHO the two data points that give the MT-215 climb rate "nose" at low altitude seem anomalous and if we utilize engine power/altitude charts for the DB605A and reverse engineer the data for the 24.5 m/s "nose" then this does not add up to the poorer performance higher up and vice versa. I think this is actually the strongest indication that things are not right with the MT-215 data: The engine performance and the climb curve simply do not match.

It is also interesting to see the wide spread in the historical data. Seeing how different the curves look it is difficult to reconcile them all with the same engine: While some (e.g. the simulation and Grundausfuhrung) show the curved characteristic of the DB605A, others are suspiciously straight. In addition, the Grundausfuhrung curve also looks strange with it's curved shape above 6 Km.

I'm happy to note that the C++ simulation falls in the middle though.

Me109climbcomp2.bmp

[Crump]

 

Speaking of accurate modelling, I noticed you posted a diagram on the Me109G2 turn capability in the paralell acceleration thread which shows the Me109G2 having a best turn rate of around 22.5 deg/s. I'm really impressed by this since this translates to 16 seconds to complete a full circle which seems to "disparage" historical Me109G2 data at 20 -21 seconds and even beats the Spitfire which took 17 seconds

 

 

You have to convert it to the same conditions of flight and you will find it agrees very well with the measured data.    No offense meant and I do not care if you ignore me or not.

[Kurfurst]

Getting back on a more constructive track:

 

Cpt_Branko, thanks for your input. While I do agree that the climb time is an important metric, I still think the climb rate versus altitude chart is a better tool for the comparisons we are making here related to the flight model:

 

I added the historical climb rate data I have to the climb chart (updated attachement) and also did an integration to derive the climb time estimates. The intersting thing when comparing the historical climb times was as you pointed out that the time to altitude was close in some cases like the Rechlin and MT-215 data. However, while this could be used to say they are equal if the target is an intercept at say 6 Km altitude I think it misses the point about the difference in performance in different height bands:

 

The MT-215 data indicates an enormous advantage up to 3 Km altitude after which the Rechlin 109G climbes better up to 6 Km which just about evens out the climb time results. However, if we choose either model this will impact at which altitude the ingame Me109G2 will perform better: If the Finnish data is used the result will be vastly superior climb performance at 0-3 Km and poorer performance at intermediate altitudes up to 6Km, i.e. it will affect the way the aircraft performs and is used at those altitudes even if the climb rate to 5 Km altitude is about the same.

 

Again, IMHO the two data points that give the MT-215 climb rate "nose" at low altitude seem anomalous and if we utilize engine power/altitude charts for the DB605A and reverse engineer the data for the 24.5 m/s "nose" then this does not add up to the poorer performance higher up and vice versa. I think this is actually the strongest indication that things are not right with the MT-215 data: The engine performance and the climb curve simply do not match.

 

It is also interesting to see the wide spread in the historical data. Seeing how different the curves look it is difficult to reconcile them all with the same engine: While some (e.g. the simulation and Grundausfuhrung) show the curved characteristic of the DB605A, others are suspiciously straight. In addition, the Grundausfuhrung curve also looks strange with it's curved shape above 6 Km.

 

I'm happy to note that the C++ simulation falls in the middle though.

 

*sigh*

 

As explained numerous times, the peak in the 0-2500m climb regime for the MT 215 is a result of the radiator settings during the Finnish trials. The Finns started the climb with closed radiators (presumably left them on automatic control) and the radiators opened fully at ca. 2500 m as the coolant temperatures rose, hence why the peak in the 0-2500 m region: the aircraft had much less drag due to the radiator flaps being in the closed position.

 

In some of the other tests, ie. the Messerschmitt testbed aircraft G-1 (with many non standard fittings, ie. non retractable tailwheel, tall tail, assymetric radiators) WNr 14 026 trials the radiators are noted to be in FULLY OPEN position up to rated altitude, hence more drag, and lower climb rates than compared to other tests. The "avarage flight model 109G" that is shown is again basically the duplicate results of the WNr 14 026 tesbed trials of August 1942 as a bit examination of the source document reveals.

 

Rechlin did flight trials, it seems its early 1943, the conditions and aircraft is unknown, though if we go by German standard its likely that the radiator flaps were 1/2-2/3 open during the climb.

 

Basically the MT 215 trials agree well with Rechlins numbers and the Soviet results with G-2 WNr 14 513 except for the 0-2 km peak, which as explained is almost certainly down to the closed radiators of MT 215 at low level climb. In short, close the radiators of 109G-1/2 and you will get around 24 m/sec climb rate, open them fully and you will get more around 17 m/sec.

[LLv34_Flanker]

S!

 

 The Finnish tests were done at a bit higher speed, around 300km/h, thus the cooling effect was better and radiator flaps stayed closed/almost up to 2500m. Normal climb speed seems to be 250km/h or so. Just a pointer.

[Crump]

 

The engine performance and the climb curve simply do not match.

 

 

 

Comparing an engine chart representing level performance will not match engine performance in a climb.

 

I would look at your propeller efficiency too.  The standard .8 efficiency tends to underestimate climb performance.

 

In the climb, density remains a function of altitude, diameter is fixed by design, and only rpm increases lowering our Coefficient of Power.  The general effect on efficiency can be seen below:

 

 

Here is the whole report from the NACA on a method of estimating propeller efficiency.  You might find it useful to fine-tune your estimations to better agree with measured results.

 

Good Luck and have fun!!

 

 

 

 

Propeller efficiency charts.pdf

[Holtzauge]

 

*sigh*

 

As explained numerous times, the peak in the 0-2500m climb regime for the MT 215 is a result of the radiator settings during the Finnish trials. The Finns started the climb with closed radiators (presumably left them on automatic control) and the radiators opened fully at ca. 2500 m as the coolant temperatures rose, hence why the peak in the 0-2500 m region: the aircraft had much less drag due to the radiator flaps being in the closed position.

 

In some of the other tests, ie. the Messerschmitt testbed aircraft G-1 (with many non standard fittings, ie. non retractable tailwheel, tall tail, assymetric radiators) WNr 14 026 trials the radiators are noted to be in FULLY OPEN position up to rated altitude, hence more drag, and lower climb rates than compared to other tests. The "avarage flight model 109G" that is shown is again basically the duplicate results of the WNr 14 026 tesbed trials of August 1942 as a bit examination of the source document reveals.

 

Rechlin did flight trials, it seems its early 1943, the conditions and aircraft is unknown, though if we go by German standard its likely that the radiator flaps were 1/2-2/3 open during the climb.

 

Basically the MT 215 trials agree well with Rechlins numbers and the Soviet results with G-2 WNr 14 513 except for the 0-2 km peak, which as explained is almost certainly down to the closed radiators of MT 215 at low level climb. In short, close the radiators of 109G-1/2 and you will get around 24 m/sec climb rate, open them fully and you will get more around 17 m/sec.

I agree: *sigh*

 

You have doggedly argued that the radiator being closed in the MT-215 test explains the 24.6 m/s climb rate. No data to support this theory has been put forward other than verbal arguments.

 

The first problem with this is that correlation is not the same as causation: If the MT-215 test was flown by a spindly Finn while the German tests were flown by a hefty Bavarian does that mean that we draw the conclusion that pilot weight is the reason for the difference in climb rate? No. Both the pilot weight and radiator flap position do of course influence climb rates but by how much? Here you take a leap of faith though an conclude that since the German tests were done with open radiator and the Finnish with closed so is the case.

 

The second is as said before that apart from verbal arguments there is no evidence to support your claim that the 7.6 m/s spread in the historical data can be attributed to the radiator being open in the low end results while closed in the high end results.

 

There is however, as I already pointed out in previous posts, both climb and speed measurement data on the Me109G that indicates that the impact on climb speed from varying the radiator flap positions is around 1 m/s not 7.6 m/s. This climb measurement data can be found in Messserschmitt report 8 18 205 001 which can be found at the WW2aircraftperformance.org.

 

There is also speed data where the difference between fully open and fully closed is around 50 Km/h, i.e. 560-505=55 km/h, 605-560=45 Km/h and 517-470=47 Km/h (From different Messerschmitt reports).

 

Lets assume the worst case scenario with a 55 Km/h speed reduction measured at 2050 m altitude (From Messerschmitt report V-109-19-L-42 which also can be found at the WW2aircraftperformance.org.) This speed difference can be used to calculate the delta in Cdo between open and closed radiator:

 

Cdo_open= Cdo_closed* (v_closed/V_open)**3

 

Which gives about a 36% increase in Cdo.

 

Entering this into the C++ model, the top speed of the Me109G2 at 2 Km is reduced from 571 Km/h to 515 Km/h i.e. 56 Km/h which is pretty close to 55 Km/h......

 

So now we established how large the impact is on drag let's see how much this affects the climbrate:

 

Turns out it is reduced from 18.68 to 17.10 m/s i.e by 1.58 m/s.

 

So to conclude, indications are that the climb rate difference between fully open and fully closed radiator on the Me109G2 is in the order of 1 to 1.5 m/s not 7.6 m/s.

[Holtzauge]

S!

 

 The Finnish tests were done at a bit higher speed, around 300km/h, thus the cooling effect was better and radiator flaps stayed closed/almost up to 2500m. Normal climb speed seems to be 250km/h or so. Just a pointer.

 

Yes, the climb speed will of course have an impact but I think optimal climb speed at low alt was closer to 270-280 Km/h and adding 20 Km/h would most likely not affect the rate to much. Concerning the radiator flaps I think I answered that above!

[Holtzauge]

Comparing an engine chart representing level performance will not match engine performance in a climb.

 

I would look at your propeller efficiency too.  The standard .8 efficiency tends to underestimate climb performance.

 

In the climb, density remains a function of altitude, diameter is fixed by design, and only rpm increases lowering our Coefficient of Power.  The general effect on efficiency can be seen below:

 

 

Here is the whole report from the NACA on a method of estimating propeller efficiency.  You might find it useful to fine-tune your estimations to better agree with measured results.

 

Good Luck and have fun!!

 

No offence Crumpp, but I think it would be in everyones interest if you stick to figuring out the F=m*a issue in the other thread!

 

Good luck and have fun!

[Cpt_Branko]

The impact of radiator opening is questionable. I noticed on some Spitfire IX tests a difference in approx 10-15% (pretty significant), in some planes less - who knows, in some planes maybe more. Treating radiator open/closed as a simple drag increase/decrease in simulation could bring about small inaccuracies since they have a tendency to produce some thrust which is normally bundled in with drag coefficient.

 

Of course, there is some spread in data. The MTT calculated data is based on calculation (it was just about to enter service). The tests along the line - the airframe saw some revisions. Of course not every test is going to be the same, not all machines were the same, so especially tests where single machines were tested might have some special factors there, but iirc Erla and Reichlin trials were done on a batch of airframes. Again no two factories produced exactly the same fighter.

 

Determining exactly the performance of warbirds is a minefield for a host of reasons, starting with different bits of data not agreeing with each other 1:1. However I take 4.4 minute to 5km to be as a representative "middle" figure when the fighter is used according to paper procedure (radiator position in climb, etc). Still, it's always nice to see people trying to write simulations, getting within 5-10% is definitely no small feat

[Holtzauge]

Yes, as was discussed before climb time to a certain altitude is an important metric. I also agree that it also goes some way to even out climb rate anomalies if the general assumptions are right but getting the climb rates right for altitude intervals is also important: If there are unrealistically large effects on climb rate from manipulating the radiator flaps then this could be used as an exploit and to me an exploit is a dirty word in the flight sim vocabulary.

Concerning the impact of the radiators flaps, I did the calculation of the impact on climb rate based on the high speed testing simply to show that there was additional evidence pointing towards the 1 m/s delta in climb rate between open and closed radiator and I guess we could have a discussion about how close that estimate is. I think it is close enough but frankly I don't think this so important: Don't forget that the main evidence is not this estimate but an actual test report from Messerschmitt that has curves showing exactly what we are looking for, namely the impact on climb rate for a G-model Me109. This explicitly states that the delta is around 1 m/s in climb rate. As evidence goes I think this is pretty conclusive but maybe that's just me......

Anyway, the developers have now received input from a number of different angles and people so they should now have some additional useful input on how to model this. My observation was simply that since you seem to see the fantastic acceleration rates you do in BOS, then reverse engineering these into climb rates results in a climb rate close to 24-25 m/s at 2000 m altitude which IMHO is unrealistic even in the reduced temperatures you are seeing in the Stalingrad scenario.

So to sum up: My estimate in standard atmospheric conditions (15 deg C) is as outlined in posts above a climb rate of around 18.7 m/s at 2000 m altitude. The delta in engine power going from this to the -15 deg C which the developers seem to say is the ambient temperature modelled, would according to my estimate (as outlined in the closed G2 acceleration thread) be around 10% added power. Assuming a standard atmospheric model going to -15 deg C would increase the density from around 1.2255 to about 1.3685 Kg/m**2. With these assumptions my input based on the C++ simulations is that this would increase the climb rate from circa 18.7 m/s to around 21.03 m/s which is respectable but significantly lower than you see today.

Anyway, I wanted to get all that in as an insurance in the unlikely event that this thread will also turn south

[Crump]

 

quasi science posts

 

MMM....Guess my quasi-science was enough for a graduate level degree and a laundry list of FAA certifications.

 

Seriously, the basic formula for climb rate is:

 

ROC = Velocity [( Thrust Available - Thrust required)/Weight]

 

In other words, it is heavily dependent upon the propeller efficiency and velocity.  Your graph shows us neither one nor do you list the formula's you are using. 

 

Instead, we are to trust your home cooked computer program which does not agree very well with the flight tested data is somehow the correct analysis making the German, Finnish, and Russian engineers wrong in their analysis.  Do you think that makes sense and we should jump on board screaming they are wrong and you are right?

Where is the evidence you are right?

[III/JG2Gustav05]

 

 The effects of airframe condition and a retractable tail wheel on climb are marginal.

 

it's not true. look at this report.http://www.wwiiaircr...-18-205-001.pdf

[Holtzauge]

That report is about the effect of radiator opening. Did you mean to link to other report?

[JG4_Nemesis]

By the way....why do we have a maximum of only 1.3 ata?

 

Shouldn't it be 1.42 ata?

 

Would mean loss of ~10% max. engine Performance...

Edited January 4, 2016 by I/JG27_Nemesis

[L3Pl4K]

 

 

By the way....why do we have a maximum of only 1.3 ata? Shouldn't it be 1.42 ata? Would mean loss of ~10% max. engine Performance...

early db605a has problem(1.42ata) with oil flow-> bubbles-> burning engine. Powersetting was locked to 1.3 ata.

[Dakpilot]

Initially the DB605 was restricted to 1.3 ATA due to piston failures, main bearing problems and oil supply issues, when these were sorted out, new piston design and different bearing material and improvements to oil system, except for a  lingering oil pressure issue, the restriction was lifted and 1.42 allowed, but for the time period of BoS the restriction was still in place

 

Cheers Dakpilot

[Kurfurst]

Initially the DB605 was restricted to 1.3 ATA due to piston failures, main bearing problems and oil supply issues, when these were sorted out, new piston design and different bearing material and improvements to oil system, except for a  lingering oil pressure issue, the restriction was lifted and 1.42 allowed, but for the time period of BoS the restriction was still in place

 

Cheers Dakpilot

 

Exactly. 1,42 ata rating for 605A-1 did not receive final clearanceon LW level until October 1943, though internally DB have cleared in June 1943 already. However it appears from several documents that the 1,42ata rating was briefly cleared for use at the end of 1942, the timeframe of BoS, but must have been recalled in the spring 1943.

 

The piston design does not appear to have significant effect on it as it only effected pre-production series and was solved practically immidiately, however lubrication problems continued to plague the 605 until an perfected oil de-aerator was fitted.

:

Don't forget that the main evidence is not this estimate but an actual test report from Messerschmitt that has curves showing exactly what we are looking for, namely the impact on climb rate for a G-model Me109. This explicitly states that the delta is around 1 m/s in climb rate. As evidence goes I think this is pretty conclusive but maybe that's just me......

 

No such document exist, in fact the very document you refer to is misinterpreted and misrepresented by yourself as you admit. In reality, the Messerschmitt trial in question shows that there is already 1-1.5 m/sec climb rate difference between fully open and half open radiator flaps, never mind when the system is fully or almost closed.

[JtD]

If you have more than the below DB605AS document to show the impact on climb, please share it. I hate to do theoretical debates, if the info exists in black and white.

 

To non-German speakers:

 

This is a test comparing three different radiator settings, rear flap at 360, 265 and 180mm. The first is fully open, the latter half open, and they tested one in between to see if a reduced maximum opening would still give sufficient cooling while also improving performance.

 

Basically they found that reducing maximum opening by 100mm improves climb performance by somewhere around 1m/s - 1.5m/s, without significantly reducing cooling capability. Setting it to half closed, they ended up with a climb rate improved by about 1.5-2m/s, but cooling capability was reduced to unacceptable levels.

 

http://www.wwiiaircraftperformance.org/me109/VB-8-18-205-001.pdf

 

Charts are on the last page, top is rate of climb, middle is time to climb, bottom is cooler suitability (has to stay above 1 to be suitable). Would be interesting to have more than a DB605AS document.

Edited January 11, 2016 by JtD

[Kurfurst]

If you have more than the below DB605AS document to show the impact on climb, please share it. I had to do theoretical debates, if the info exists in black and white.

 

To non-German speakers:

 

This is a test comparing three different radiator settings, rear flap at 360, 265 and 180mm. The first is fully open, the latter half open, and they tested one in between to see if a reduced maximum opening would still give sufficient cooling while also improving performance.

 

Basically they found that reducing maximum opening by 100mm improves climb performance by somewhere around 1m/s - 1.5m/s, without significantly reducing cooling capability. Setting it to half closed, they ended up with a climb rate improved by about 1.5-2m/s, but cooling capability was reduced to unacceptable levels.

 

http://www.wwiiaircraftperformance.org/me109/VB-8-18-205-001.pdf

 

Chart are on the last page, top is rate of climb, middle is time to climb, bottom is cooler suitability (has to stay above 1 to be suitable). Would be interesting to have more than a DB605AS document.

 

Indeed.

 

Note however that even with "half zu" (half open, 180 mm opening) the Gütegradf is above 1 (meaning suitable) up to around 2500 m altitude, coincidentally the same altitude was noted by the Finns at which the radiator flaps would open during climb and reduce climb rate. The finns however did climb at a little higher speed (ca 300 kph) so as a result cooling flow was improved compared to the suggested climb speed (ca 270).

 

Curiously the greatest gain from going from 360 mm to 180 mm radiator opening occurs at the rated altitude of ca 6000, where the difference is ca 2,5 m/sec. This is the difference between fully open and half open radiator flaps.

 

This is nothing new in this, radiator flaps are a source of considerable drag, for example Spitfire trials also showed some 3,5 m/sec improvement in climb when going from 'open' to 'closed' radiator flaps.

[Holtzauge]

MT 215 climb results are, for example, perfeclty normal and fit very well to other known tests, if the testing prototocol deviation in that tests (closed radiators at the beginning of the climb) are taken into account. The radiator position of the 109 had a strong influence on drag and rate of climb.”

No, the MT-215 climb test is not perfectly normal. There are no other climb test results that come close to 24 m/s climb rate at 2 Km for the Me-109G2. If there were I’m sure you would have posted them on your website years ago. You also conveniently avoid mentioning what the Finns themselves said about the MT-215 tests: they classed the speed test as reliable but the climb test as unreliable due to no special test equipment used to gauge the climb rate. Ever stopped to consider that the 24 m/s may simply be a measurement error instead of arguing the result as “representative”? However, seeing your penchant for high end performance outliers when it comes to the Me-109 I’m not surprised….

 

Basically the MT 215 trials agree well with Rechlins numbers and the Soviet results with G-2 WNr 14 513 except for the 0-2 km peak, which as explained is almost certainly down to the closed radiators of MT 215 at low level climb. In short, close the radiators of 109G-1/2 and you will get around 24 m/sec climb rate, open them fully and you will get more around 17 m/sec.

No, there is absolutely nothing indicating this. What you call certainty looks more like maskirovka. In post #26 there is a calculation for the effect of going to a FULLY closed radiator also based on Messerschmitt data and this increases the climb rate when going from the half open value of 1 m/s to 1.58 m/s when fully closed at 2 Km altitude which is nowhere close to the 7.5 m/s you claim. Do you honestly believe going from fully open to half closed radiator gives you 1 m/s in climb and then when you go from half closed to fully closed you get the other 6.5 m/s you’re missing or is that just the lawyer talking?