Use the lowest output voltage available on your square wave source that is still clearly visible on the scope. Square waves stress an amplifier much more than audio, so they need to be tested at a low level to be applicable to playback behavior.
No more toasted outputs
And the choir says AMEN!
With low impedance loads you get more load distortion. With high impedance loads you get more voltage distortion. I think the available power BJTs are more linear with 8R loads than 4R, and can output more power that way as well, and be faster, though at the cost of higher rails.
I think the D4xH11 are better suited to 4R loads, and they are faster than the MJL's as well. They would also be good on a bridged amp with an 8R load.
Last edited:
Hi Guys,
I would like to designate on set of boards to try with high rails, say +-90V. I have a few output transistors on hand and would like suggestions as to which ones would be the best suited for the task. Here's a list of the ones that I have at least 10 pair.
MJL4281/4302
MJL4281G/4302G
MJL3281/1302
MJL21194/21193
MJL21194G/21193G
NJW0281G/0302
NJW21194G/21193G
I would also like a suggestion as to which of the IPS might be the best choice for high rail operation.
Thanks, Terry
I would like to designate on set of boards to try with high rails, say +-90V. I have a few output transistors on hand and would like suggestions as to which ones would be the best suited for the task. Here's a list of the ones that I have at least 10 pair.
MJL4281/4302
MJL4281G/4302G
MJL3281/1302
MJL21194/21193
MJL21194G/21193G
NJW0281G/0302
NJW21194G/21193G
I would also like a suggestion as to which of the IPS might be the best choice for high rail operation.
Thanks, Terry
i didn't like this at all...
In any case we should note that there are some lurking discover a mistake to some grinning.
Too bad things happen to such a situation in a forum that one nighters to offer something.
Terrible persuaded to see some to bet on the failure of others.🙁
Thimios
In any case we should note that there are some lurking discover a mistake to some grinning.
Too bad things happen to such a situation in a forum that one nighters to offer something.
Terrible persuaded to see some to bet on the failure of others.🙁
Thimios
I'm not sure I understand you. Can you rephrase? You make it sound like someone made a deliberate error at the expense of others. Are we missing something else?
Perhaps some of the remarks about the simple mistake mixing Mhz and Khz would explain those comments. I know that was a simple error and not intended to confuse.
Maybe it was all this prodding that any properly operating amp should be able to do 20Vpp at 200k hz without smoking the zobel. In the last scope test I did, I set the generator so that I saw 10vac on the output at 1Khz. Then leaving the volume pot alone I began graduating the frequency upwards and simply changed the time dial on the scope. The relative height of the square wave on the screen stayed the same but the voltage reading on the DMM began to drop when I reached the higher frequencies. At 100khz I tried raising the level some while watching the DMM and then noticed smoke rising from the zobel. I wouldn't be surprised if there were no zobel resistor to reveal that, an output would be the next thing damaged.
An amp should be able to do 100KHz full power without smoking, because that's the fastest signal you can get out of a standard soundcard these days. Of course if you have a LP filter at the input, it's not quite as severe.
One benefit to having very fast amps is that the C in the output RC can be small, so the R is less likely to burn. In any case I'd use carbon comp for the zobel resistors for good surge tolerance in case the amp gets "excited" around CB radios or whatever. Multiple in parallel would be even better for low inductance.
One benefit to having very fast amps is that the C in the output RC can be small, so the R is less likely to burn. In any case I'd use carbon comp for the zobel resistors for good surge tolerance in case the amp gets "excited" around CB radios or whatever. Multiple in parallel would be even better for low inductance.
Yes, but a 100kHz sine is a very different animal from a 100kHz square and some seem to have been seduced into feeding their amps squares. The much higher harmonic content is way harder on the amp.
That's why I suggested to do the square wave test at a very low level. 100KHz is the fastest signal a soundcard can produce though, which is why I consider it the worst case scenario to be safe against - it's a practical thing.
I don't know what a sound card has to do with it. I can only tell you what I have experienced along with the several warnings I have seen from folks I trust that say don't pump high level square waves through your amp at high frequencies. I suppose you must do it all the time since you continue to encourage it. Perhaps this is what thimios is referring to.
Edit; Looks like we were typing at the same time. Square waves are what we were talking about. Read thimios's post again.
Last edited:
If you understand how the amp reacts to RF, and know that it won't blow up, then you're safe. So yes, I do it all the time because I don't design amps that smoke from RFI.
It doesn't seem like you have those skills, so no, I wouldn't suggest to do any high-level high-frequency tests. But it's okay for you to do the low-level square wave test because it's applicable to you - it shows audio signal behavior and you have no need to test for RFI sensitivity.
This thread has both designers and testers. There is advanced information for the designers and safe guidelines for the testers. I guess the distinction is not very clear right now.
It doesn't seem like you have those skills, so no, I wouldn't suggest to do any high-level high-frequency tests. But it's okay for you to do the low-level square wave test because it's applicable to you - it shows audio signal behavior and you have no need to test for RFI sensitivity.
This thread has both designers and testers. There is advanced information for the designers and safe guidelines for the testers. I guess the distinction is not very clear right now.
I think people are mixing things up here, I certainly was. Keantoken isn't advocating that an amp should be able to reproduce a full power 100kHz square wave, just a 100kHz sine wave as that is the maximum frequency any 192kHz sample rate stuff can contain. Granted you'll be hard pushed to find any media that contains any full output content at 100kHz, but in theory it could be there.
It's clear now. He's a designer and I'm not so I should disregard his suggestions because they are for designers and not us testers.
Got it
Got it
realistically , class G or H would be the next level beyond a 4 or 5 pair
fixed rail BJT OPS.
To allow our fast slewing amps .... "riding the moving rails"
AV: Riding The (Moving) Rails: Detailing Class-G And Class-H Amplifier Topologies - Pro Sound Web
would need to be on the "fast track" !! 😀
OS
The ones highlighted in the quote are likely the best for pushing the operating voltage. As for which IPS that I'd see if OS chimes in though I'm sure they can all operate at any voltage the OPS can handle.
Thanks Jason.
I was leaning towards those but was also thinking the 21194/93 because though they are less voltage, they have higher amp rating. As for the IPS, I figured there might be one that is best for driving that type of amp. I'm thinking of pulling out my Leach Superamp and using the chassis, heatsinks and power supply and building what OS calls a "party amp". That has a 65-0-65vac 800va transformer in it so should be pretty impressive. Who knows, the CFA-XH might be the best one for that and if so, I'll be set since I have a few of those boards on order. My other two pair of OPS will likely never see a chassis as I plan to use those for an A/B setup for comparing IPS designs. 😉
Blessings, Terry
I was leaning towards those but was also thinking the 21194/93 because though they are less voltage, they have higher amp rating. As for the IPS, I figured there might be one that is best for driving that type of amp. I'm thinking of pulling out my Leach Superamp and using the chassis, heatsinks and power supply and building what OS calls a "party amp". That has a 65-0-65vac 800va transformer in it so should be pretty impressive. Who knows, the CFA-XH might be the best one for that and if so, I'll be set since I have a few of those boards on order. My other two pair of OPS will likely never see a chassis as I plan to use those for an A/B setup for comparing IPS designs. 😉
Blessings, Terry