| Edwin Dorre |
Hello,
what is the test way to test if my Aleph 4 is pushing enough watts?
I figured the following, however I first wanted to ask you guys before trying...
Take a signal generator on 100hz sinus and connect it to the Aleph. Take a 8 ohm big resistor and connect to speaker and connect a scope to this resistor.
Now pump up the amplitude of the sinus till I see the output sinus clip. Read the +/- voltage of the sinus and do some calculations.
I=voltage-top-top / 8
W=voltage-top-top * I
Is this correct?
Edwin |
|
|
| jewilson |
The power calculation is (Voltage Squared) / the load.
Example, if an amp swings 40 volts across and 8 ohm resistor
the power is 40*40=1600/8 =200 watts
If you have an amp that make this much power and you only
have a 100 watt resistor you still use it. The resistor must be
placed in a can of oil to dissipate the heat.
Rise the input voltage till the peaks just start to flatten and stop.
The measurement should made with a scope and meter. |
|
|
| grataku |
Edwin,
what jewilson said is essentially correct, HOWEVER, since to "swing" means ALOT of things, he should have probably mention "root-mean-square, RMS, output voltage" that is Vpp*(1/(2*sqrt(2)) (Vpp=volt peak-to-peak). So what you get at the end is the very legit watt RMS figure using either the V^2/R or RI^2 formula.
Modern digital scopes ($$$$) provide direct RMS voltage measurements. So unless you have such capabilities you have to do the conversion. Usually, 'normal' (ie your regular $100-type) digital multimeters have a very narrow frequency range and are not reliable much above the mains frequency. |
|
|
| GRollins |
I use a Fluke 8060A (granted, more than $100). It gives true RMS readings, which simplifies things tremendously. I hook up one of those big gold Dale 8 ohm load resistors and take a direct reading.
To establish clipping, I use a triangle wave since it's a little easier to see when the top and/or bottom flatten. I then switch back over to sine. The voltage reading you get off a triangle wave will not be directly applicable to your usual RMS calculation.
Grey |
|
|
| swede |
So, where do you guys get thoose hefty 8 ohm resistors?
//magnus |
|
|
| jackinnj |
| Calories per minute equals watts. So if you heat a resistor in a thermos containing one litre of water with your amp and measure the temperature change over a few minutes you can find watts |
|
|
| jewilson |
| Dale is one good source for high powered resistors. |
|
|
| PassFan |
| Don't know if it's right or wrong, but I always measure my rms output voltage into the actual impedance (i.e. speaker) I will be driving. I always figured the amp would react differently under a real world load versus a resistor. A quick check on the common houshold circuit will tell you if the meter reads true rms. If you see what your supposed to see (120v or thereabouts) your looking at true rms.;) |
|
|
| jewilson |
You can do all types of damage driving a amp to full power across the speaker. At some frequencies, the speaker changes impedance's and so might the amp. When the amp goes non-linear it will pulls the voice coil up on against the magnet blowing your speaker. Even a under powered amp can blow high power speakers this way. Also, the speaker can fail when it bottoms
out from to much power.
The standard is to use a resistor, If you want to measure the speaker impedance over frequency that a different story. |
|
|
| wuffwaff |
Hi Edwin,
Iīve build myself an adjustable resistor for 8/4/2 ohms. It is made out of 2 heatsinks and 16 8Ohm 25w dale resistors. I can test 1 x 16 ohm, 2 x 8 ohm, 2 x 4 ohm and 1x 2 ohm.
I also use greys method of Fluke (used, about 150 ) and scope and check for clipping on an triangular wave.
For the Aleph 4 I think you could do without a heatsink if you use 4 x 16ohms 25w. This way you can test 8 and 4 ohms power.
william |
|
|
| Nisbeth |
If youīre having trouble finding large 8 ohm power resistors you can parallel 7 56 ohm or 15 120 ohm power resistors for higher wattages.
/U.
PS: Oh, yes put them in an oil bath for cooling. When the room smells of popcorn, you know itīs time to turn it down :D |
|
|
| jackinnj |
| Too late to edit it I guess -- if you heat a litre of water 1 degree in one minute you have expended 69.7 watts of energy. (The equation is 1 cal/minute = 0.0697 watts). You're probably better off using a good, levelled, random noise source than a sine generator. |
|
|
| JBL |
Puting the resistor in a bucket of pure water work well also if you don't like oil.
Just make sure it'S pure |
|
|
| peranders |
When I tested my QRO-amp I used Arcol 300 W non-inductive
http://www.elfa.se/pdf/60/06063002.pdf
http://www.elfa.se/elfa/produkter/se/20/2012534.htm
I put the res on a huge heatsink which came from a locomotive (the swedish RC locomotive type from ABB, ADTranz, Daimler-Benz, Bombardier or what they call themselves these days)! The fins were 6-7" long. This heatsink was placed in a buck of water. 600 W of continious power got the water boiling! |
|
|
| Edwin Dorre |
Cool,
I own a TrueRMS voltage meter at least it is saying so on the outside. Do I use the voltage meter to measure the drop over de resistor or the amperage in serie with the resistor? I lost the manual of this critter and I am not that smart with electronics...
Is a 100Hz sinus ok? Or can I better use an other frequency?
Thanks,
Edwin |
|
|
| wuffwaff |
Hello Edwin,
you should measure the voltage drop over the resistor.
I would go for 1kHz and then you can always test at 10/100 Hz and 10/20/50/100 kHz.
william |
|
|
| grataku |
| I would just leave all this bulls about buckets of oil or water to the local McGyver and limit the discussion to the 'humane' way of measuring powers. |
|
|
| PassFan |
Jewilson:
"You can do all types of damage driving an amp to full power across the speaker."
How can I hear any sound if I don't hook it up to a speaker? Is your reference to, "full power", peak before clipping, or clipping at full output.
Yes, I understand the speaker must be matched to the amp and pushing an amp into clipping will overheat the voice coil in the speaker and you don't have to be at full power to blow tweeters.
I appreciate your concern and the warning but I think we have to be a little more specific in light of the fact that some, not all, of the new guys don't know about some of this stuff yet.
As for my unorthodox method, it is just to get a general idea of how much I am really pushing. Believe me it isn't easy cause the meter won't stand still. Thanks:) |
|
|
| Tube Addict |
On the web you can find several 'shareware' scope-software for measurments which use your pc's soundcard (all soundblaster compatibels will do the job)
Anyone have experience with it?? Have not tested it myself because of possible burnouts of my soundcard...
Greetings,
Tube Addict |
|
|
| roddyama |
Hi William,
| quote: | | I would go for 1kHz and then you can always test at 10/100 Hz and 10/20/50/100 kHz. |
Most typical DVOM (Digital Volt-Ohm Meter) will measure AC voltages greater then 1kHz at an attenuated value. I have the Fluke 8060A ($400 12 years ago) and it's good to 20kHz. You need a scope or a more expensive desk DVOM to make the high frequency measurements.
Rodd Yamashita |
|
|
| wuffwaff |
Hi Rodd,
I thought Edwin said he had a true rms voltmeter I didnīt think of any frequency range. Iīve bought an old fluke (donīt know the type right now) wich seems to have no problem with 200kHz or more.
william |
|
|
| Edwin Dorre |
| quote: | Originally posted by wuffwaff
Hi Rodd,
I thought Edwin said he had a true rms voltmeter I didnīt think of any frequency range. Iīve bought an old fluke (donīt know the type right now) wich seems to have no problem with 200kHz or more.
william |
I found my manual of my multimeter. It will measure up to 1Khz in AC-Current mode so I have to stay below this...
Ok, I am of to find a good resistor. Mine only goes to 30W, so I will give it a quick test try with two papst 8x8 fans on it...
Edwin |
|
|
| PassFan |
| You had better be careful as you don't want to blow it while your driving your amp, it could cause damage. It's not going to give you a warning or put it's hand up when it happens either. Do it quick if you must. Another idea is a dryer element. The wire they use is a resistive element type of wire and you can cut it and set it up for whatever resistance you need. Remember, Nelson used light bulbs in an amp though the resistance on a bulb changes when it heats up. Using the wattage of the bulb and the voltage output you can calculate the resistance using ohms law. Good Luck |
|
|
| Edwin Dorre |
Hello,
just did my testing. With a 8 dale 50W 1ohm resistors and a dale 160W 5 ohm resistor...
With 8 ohm, it tops out at 25.4V RMS = +/- 80W
With 5 ohm, it tops out at 15.32V RMS = +/- 46 W
hmm.... Sounds a bit too low doesn't it?
When I change channel, frequency or resistance, sometimes it tops out first and sometimes it bottoms out first... Can someone tell me why?
The current running through my 1.5ohm resistors of the IRFP240's run between 0.565V and 0.588V, can this be why? How can I raise this?
Edwin |
|
|
| Nelson Pass |
I don't feel like doing any work today, so I'll just put
in my 2 cents.
1) Very few RMS meters have a decent bandwidth, and if
you don't have a bandwidth spec, then measure at 100
Hz or so.
1A) Radio Shack makes fine meters if you observe this
limitation. I keep about 10 around at all times.
2) As long as you are running a sine wave, RMS and averaging
meters will give you the same answer. Again, watch the
bandwidth.
3) You can get away with driving load resistors way over
their rating if you do it fairly quickly. |
|
|
| wuffwaff |
Hello Edwin,
sounds good to me.
In both cases youīve got about 3.1A effective current wich means about 4.4A peak. This fits the 2.3A bias I calculate from your voltages/resistors.
You can change this value by raising the value of R19 (Aleph5). Set this with a trimmer and insert the right resistor afterwards.
Depending on the setting of the current source gain you get different values of maximum current. When using 50% you`ll get a maximum of twice the bias current.
When you have less current gain then the max. current will be lower than twice the bias current, when youīve got more it will be more.
I have to set my Aleph 5 to about 70% current gain to get the specified 90watts/4 Ohms.
Iīve calculated this value for the whole Aleph range looking for max output current specified in the manuals.
For example Aleph5 2A bias and 8A max so 4xthe bias current would mean a current gain setting of 75%
The others (from 1.2 to 60) hoover around 60 to 65%
First check your current gain setting and your bias and youīll see that it fits.
Normally the amp will always clip on the positive side when the "negative clipper" is set for a generous amount of current. If you want the negative side to be able to sink more amps youīll have to change this by lowering R16.
william |
|
|
| jackinnj |
If you look at the pdf's for the Analog Devices AD536 true RMS converters (there are several flavors, I have a ton of AD737's), but the 536 is the device which Charles Hansen uses in his various designs in AudioXpress, you'll note that a capacitor is necessary for integrating, it is root MEAN square, after all. A large cap is needed for low frequencies and vice versa.
"They laughed when I said: "Use the calorimetric method"". Too many p-chem labs, I guess. (Oh, did I forget to mention that Linear Tech once made a true RMS converter which used a heating element. I've never been able to aquire one, however.) |
|
|
| Edwin Dorre |
I am getting confused!
In the service manual of the Aleph 4 is written: 100W at 1% at 8 ohm !
Clipping at 80W is hardly 1% is should think. Also there is written bias is slightly higher than 3A (the same is in the Aleph 2 manual). But one runs 2.4A (mine 2.3A) and the other about 3A.
Why do I get 3.1A when the max is about 2.4A at 47% = 4.7A or so...
I am a bit of a novice; I bought a 800page book; Learning Electronics and Electricity just to learn how the Aleph is working. I read a couple of chapters of the passive components but I am not getting a grip of the interaction of the active and the passive. And I am starting to get puzzled about active feedback loops and stuff...
So my question is why is a Aleph 2 running 160W at 4 ohm and 100W at 8 ohm with a current gain of 46% and my Aleph 4 running 46W at 5 ohm and 80W at 8 ohm with a current gain of 50% .
(0.07833 * 1000) / (0.25 * 619) = 0.50
When my calculations are ok the Aleph 5 runs at 1.5A and has a current gain of 78% ...
Edwin
__________________________
Meten is weten ... toch? |
|
|
| wuffwaff |
Hello Edwin,
donīt get confused:)
Thereīs a difference between peak current and rms current. The max peak current at 50% current gain will be twice the bias current.
Max power will be peak current diveded by sqrt2 times the max eff. voltage (this will be reached at just one load)
The bias of the Aleph4 is aprox. 2.6A not just over three as is in the service manual. You can calculate this with the 0,6V over the 1.5ohm resistors or the 250W channel at 96volts supply.
2.6A will give 5.2A peak and 3,67A rms giving 3,67x3,67x8 = 108w into 8 ohm
This fits:D
Your bias is a bit down so power is also down.
When power is limited by the current it will halve as the load halves so from 80/8 in 40/4ohms
The Aleph 5 has about 2A bias.
Please read the patent again and again, it really helps!
william |
|
|
| Edwin Dorre |
Ok, how do you calculate the 2.6A ? When I do;
I = V / R = 0.6 / 1.5 = 0.4 * 6 = 2.4A ...
I still do not understand how Aleph 2 should be able to pull 160W at 4 ohm (ok, I do not understand a lot of things).
0.5 / 1 * 6 = 3A . 47% gives about 3A to 5.9A . For 160W in 4ohm you need about 6.3A rms .
5.9A gives about 5.9 * 0.707 = 4.17A rms = about 70 watt!
So you need a max bias of 6.3A * sqrt2 = 8.9A to pull 160W at 4ohm ...
Some how I am starting to think the services schema's of Nelson are not in sync with the specs... ;-)
Edwin
_________________________________
Please note I am a novice at electronics so be gentle with me! |
|
|
| wuffwaff |
Edwin,
2.4A is what you get when you use the ca.0.6V value, 2.6A is when you take 500watts/96V/2.
The bias value can only be an estimated value cause it depends on Vgs and the hfe of the biasing transistor so it will vary a bit depending on what youīve used.
Your calculation for the Aleph2 is right, but where did you get the 47% from?I think somewhere was mentioned that the formula to calculate the current gain is only a aproximation and that itīs much easier to measure.
The Aleph2 will do 200w@4ohm with 3.3A bias and 67% cs gain for 10A peak.
Just try to get your bias up to 2.5-2.7A and set the current gain as you like it best and enjoy the sound, donīt worry too much about the specs :cool:
william |
|
|
| Edwin Dorre |
Hello William,
first of all I like to thank you for all the time you took to answer my questions. I could have build the Aleph without knowing how it works (I had enough building, soldering, component reconizing,- skills), but I have to know how it works...
I will read the patent a couple of times more with my electronic books and your comments...
I have used some other serie resistors like the 220 ohm instead of the 221 ohm . I used some E96 for the 392 ohm. I think I used a 4.70 for the 4.75 and a 220K for the 221K ... I will swap some around...
This can account for the lower bias, can it?
Edwin |
|
|
| Edwin Dorre |
Hi, me again,
two fast question, can I raise R19 (aleph 5) or R26 (aleph 4) without any trouble to increase the bias?
Or do I need to do something with R23 (aleph 4) as well?
Edwin |
|
|
| wuffwaff |
Hello Edwin,
I did the same using 220R, 390R, 4k7 etc. This wonīt harm and mostly the diff isnīt even 1%.
This normally canīt account for the lower bias. Your bias isnīt that far of so maybe itīll suffice to raise the value of R26. If this is not enough you can look for a transistor with a lower hfe (can be the same type but a BC550C would work too).
In the last instance youīll have to change the source resistors to 1ohm or so.
Normally you wonīt have to change R23. It only sees to it that the amp wonīt sink too much current. If you normally have clipping on the negative side before the positive clips youīll have to lower the value a bit.
william (he who has just received his black gate caps:) ) |
|
|
| Edwin Dorre |
| quote: | Originally posted by wuffwaff
Your bias isnīt that far of so maybe itīll suffice to raise the value of R26. |
How much can I raise R26 to keep in the green zone? I Put a 10K in serie which did almost nothing, I put 60K (actually two 120K in parallel) in serie which gave me the 0.6V over the 1.5 ohm resistors. I tried 120K in serie; for a total of 240K which raised it a lot (around .710 but I did not dare to measure long), but it kind of scared me so I switched it back off...
Is 240K still a valid value for R26?
Edwin |
|
|
| wuffwaff |
Hello Edwin,
the only reason itīs there is to reduce the dc bias. You can leave it out without any danger. At 240k it wonīt be doing very much anyway (just test it!):)
william |
|
|
| Edwin Dorre |
| quote: | Originally posted by wuffwaff
Hello Edwin,
the only reason itīs there is to reduce the dc bias. You can leave it out without any danger. At 240k it wonīt be doing very much anyway (just test it!):)
william |
With no resistor I have over 2.8A .
With 240k I have about 2.688A of current.
With 190K I have about 2.4A of current.
With 120K I have about 2.3A of current.
So I probably put in a 240k just to be sure it is still limited somehow.
I am still curious over why I need to do this adjustments to get the current up to a 'normal' level!
Edwin
P.s. Just had a big bang *spark* which shook the hell out of me, still do not know what happened but everything is still working! I was with my back to the amp, my girlfriend has seen the spark. I am going back to 20VA trafo instead of these 1000VA ones! I think my turning around I touch a wire on the floor which must have moved another wire which touched another wire and boom... can not find a black spot... Tomorrow I am going back to the amp with new courage... |
|
|
| wuffwaff |
Edwin,
your bias was at normal level! I think if you would try a few different transistors for the current source you would see changes in bias too.
The current trough the transistor + Vbe defines the working point.
william |
|
|
| Edwin Dorre |
| quote: | Originally posted by wuffwaff
Edwin,
your bias was at normal level! I think if you would try a few different transistors for the current source you would see changes in bias too.
The current trough the transistor + Vbe defines the working point.
william |
Changing to the 523ohm resistor for a other current gain upped the power to 100W and to 67W in 5 ohm. However I also changed the R26 to 240K to up the bias a bit, so I have now about 112W in 8 ohm and 81W in 5 ohm...
I am happy now! I needed a bit more juice in the 4 ohm department as I own B&W Nautilus 802 which drop a bit towards 4 ohm ...
Now I have to build the lot in a case, as all is build on a MDF board.
Edwin |
|
|
| jackinnj |
| A spark is nothing, the cap below went off like a shotgun shell -- it was from the bias circuitry of a photomultiplier HV supply -- I ran too much current through it. |
|
|
|
