Hi,
Have you measured the currents though the LTP transistors?
Have you measured the AC voltage across the NFB cap?
Have you measured the difference in voltage across the LTP bases?
Have you measured the variation of bias with temperature?
Have you measured the variation of output offset with temperature?
Could the levels of the higher harmonics be an indication of crossover distortion? What was Vre during your first and this later test?
Can you repeat the testing with more bias current?
Try Vre=15mV and Vre=20mV
Have you measured the currents though the LTP transistors?
Have you measured the AC voltage across the NFB cap?
Have you measured the difference in voltage across the LTP bases?
Have you measured the variation of bias with temperature?
Have you measured the variation of output offset with temperature?
Could the levels of the higher harmonics be an indication of crossover distortion? What was Vre during your first and this later test?
Can you repeat the testing with more bias current?
Try Vre=15mV and Vre=20mV
No. I've actually never measured current before, so I'm not sure how to do it. Across which pins of the transistors do I measure? Are there any issues I need to be aware of to do this with damaging anything?AndrewT said:
Which one is the NFB cap? I searched the acronym dictionary but none of the definitions there look likely to be what you mean!: http://www.acronymattic.com/NFB.html
No. I'll check that out. What does this measurement mean? What range should it be in?
The mV across the output resistors goes up over time as the temperature increases. While I was running the tests for an extended period, since I was working the amp very hard, the heatsinks got very hot (only able to touch them for a couple of seconds comfortably), and mV across the resistors went up to 16.0, from 8.0 which is where they were at first.
I'm not sure what "output offset" is, or how I measure it.
Sorry, I don't know what either of these questions mean, or how I test them.
I certainly can, but I'd like to try fixing the grounding issues that Arthur pointed out before I do much more testing. Another diyaudio member dropped by today and provided some suggestions on how to do this, do I'll see what impact that has.
Andrew, many thanks for your help - and sorry I'm not able to provide intelligent answers to many of your questions!
Re: Power
Any tips?...
Yeah, I was wondering about that. I can easily get the voltage measurement (I've calibrated the software against my multimeter for this purpose), but I don't know how to get the current/power. When I picked the amperes setting on my DMM and connected it to the speaker outputs whilst under load, I got a little spark and blew the fuse. So I guess that's not the right way!PHEONIX said:
Any tips?...
Some success today, after making the following changes:
I've now got THD at 1kHz-2kHz to match SC's measurements, and each channel is reasonably similar to the other (within 25% THD, compared to about 70% difference before). I can only do this at low power at the moment, because my voltage divider on the dummy load isn't attenuating enough for my sound card to handle full power. I've ordered the necessary parts to upgrade the dummy load to handle this, so I hopefully will be able to test at higher powers in the coming days. (According to SC's article, this shouldn't make any difference any way - they show no impact of power on THD, for 5-100W.)
The low frequency noise is much improved now - for instance if you compare this 1kHz spectrum to the one I showed a few days ago, you'll see 50Hz and harmonics down about 6-8dB as a result of these changes:
Next step is to reconnect the electronics to the chassis, this time via Rod Elliot's loop breaker circuit:
Hopefully that will allow me to get the safety of proper earthing, without the "dual star ground" noise I was seeing earlier.
- Use thicker wire for connecting PSU to amp
- Insulate RCA sockets from chassis
- Temporarily disconnect electronics from chassis
I've now got THD at 1kHz-2kHz to match SC's measurements, and each channel is reasonably similar to the other (within 25% THD, compared to about 70% difference before). I can only do this at low power at the moment, because my voltage divider on the dummy load isn't attenuating enough for my sound card to handle full power. I've ordered the necessary parts to upgrade the dummy load to handle this, so I hopefully will be able to test at higher powers in the coming days. (According to SC's article, this shouldn't make any difference any way - they show no impact of power on THD, for 5-100W.)
The low frequency noise is much improved now - for instance if you compare this 1kHz spectrum to the one I showed a few days ago, you'll see 50Hz and harmonics down about 6-8dB as a result of these changes:
Next step is to reconnect the electronics to the chassis, this time via Rod Elliot's loop breaker circuit:
An externally hosted image should be here but it was not working when we last tested it.
Hopefully that will allow me to get the safety of proper earthing, without the "dual star ground" noise I was seeing earlier.
is this back to front.Ray Chuck said:
Crossover distortion is worst at outputs that just exceed the ClassA output current.
At high output current the proportion of crossover distortion reduces to almost insignificance compared to the other distortions that increase with increasing output current.
AndrewT said:
Crossover distortion is a different mechanism to switching distortion. Switching distortion comes about because of minority carrier storage in the output devices causing both devices to conduct simultaneously during the crossover switching point, where as crossover distortion is caused by Gm overlapping. One is a large signal phenomena whereas the other one is a small signal phenomena. Switching distortion will be dominant for large signals(large currents) whereas crossover distortion will be dominant for small signals(low currents).
Some amps particularly those with low fT devices can't be driven hard at high levels at high frequencies for very long because the common mode conduction current can destroy the output devices
Got the non-inductive resistors today for the new dummy load/voltage divider. I'm going to use a pair of heatsinks both for heat resistance, and as the sides of the unit (so they'll work as shields too). I'm using a combination of resistors in parallel and series plus a number of different sockets connected to different parts of the series to allow me to choose different levels of attenuation by simply plugging into a different socket. Here's one of the sides, prior to do any wiring:
The big resistor is a 50W/15Ohm (which will go in parallel with another same specced on the other resistor, in order to provide 100W/7.5Ohm). The others are to handle the various different levels of voltage division.
The big resistor is a 50W/15Ohm (which will go in parallel with another same specced on the other resistor, in order to provide 100W/7.5Ohm). The others are to handle the various different levels of voltage division.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.