Thanks,got it!Thi is valid for class AB output stages.
Imagine, you measure THD at relatively high output swing - say, 20V RMS (50W @ 8 ohm).
In case your OPS is under-biased, you will have some crossover distortion, but it's pretty much "masked" by the high-level base frequency tone.
Now, measure THD at 2.8V RMS at the output (1W @ 8 ohm). Crossover distortion stays roughly the same, but the level of the base frequency is much lower - THD value increases.
In correctly-biased OPS, crossover distortion's influence on overall THD is lower, than influence of the other sources' of distortion. In this case - lower swing -> lower THD.
In my previous test THD is higher when output is lower than with a higher output.
Last edited:
Bias and clamping readjusted and amplifier was tested again.
Now an even lower distortion level measured.
It looks like there's some 50hZ noise in these. Is this actually from the amp or from the sound card?
No,no this is an artifact not hum🙂It looks like there's some 50hZ noise in these. Is this actually from the amp or from the sound card?
Last edited:
Crossover distortion is a short period glitch and thus is high frequency in nature.
You will see higher levels of 5th to 39th harmonic content when crossover is present compared to a ClassA stage which tends to have higher levels of 2nd to 4th.
D.Self shows some pics of BJT crossover effects for optimally biased, under biased and over biased. You can clearly see the spiky nature of the high frequency content of the crossover. MosFETs don't have an equivalent to BJT optimally biased. They just get lower crossover distortion the more you increase the bias current.
I'd guess that's why Borbely tells us to use 500mA, or more, of output stage bias.
You will see higher levels of 5th to 39th harmonic content when crossover is present compared to a ClassA stage which tends to have higher levels of 2nd to 4th.
D.Self shows some pics of BJT crossover effects for optimally biased, under biased and over biased. You can clearly see the spiky nature of the high frequency content of the crossover. MosFETs don't have an equivalent to BJT optimally biased. They just get lower crossover distortion the more you increase the bias current.
I'd guess that's why Borbely tells us to use 500mA, or more, of output stage bias.
try to find a way/s to attenuate the artefact.No,no this is an artifact not hum🙂
Is it a cable effect? or in the input stage of the measuring equipment?
I have tried hard but i can't always eliminate this .try to find a way/s to attenuate the artefact.
Is it a cable effect? or in the input stage of the measuring equipment?
May-be the input &output common GND is the matter.
Screened twisted two core in a balanced connection virtually gets rid of the two chassis grounds being at different voltages.
Member
Joined 2009
Paid Member
Thimios - the square waves look good. But I think you can fit some more test equipment on top of that stack you have, you only have to get a step-ladder to operate them 😀
Hi Gareth,I will try this in a next test😀Thimios - the square waves look good. But I think you can fit some more test equipment on top of that stack you have, you only have to get a step-ladder to operate them 😀
Last edited:
Ready for testing
Well, just about ready for testing. Man this thing burns up some real estate. 😱
I need to clear off my bench before I can do any more. I also want to read back through the thread to refresh my thinking before I start testing. I want to be sure I understand how to adjust everything before I go any further. I have all three IPS sections ready but I think I will start with the VFA since it seems that is the most stable.
Blessings, Terry
Well, just about ready for testing. Man this thing burns up some real estate. 😱
I need to clear off my bench before I can do any more. I also want to read back through the thread to refresh my thinking before I start testing. I want to be sure I understand how to adjust everything before I go any further. I have all three IPS sections ready but I think I will start with the VFA since it seems that is the most stable.
Blessings, Terry
Attachments
Hi AndrewT,please can you post a picture?Screened twisted two core in a balanced connection virtually gets rid of the two chassis grounds being at different voltages.
Something like this?
I have tried this in the past without success.
Attachments
Last edited:
Hi Terry,yours is a little different output board.Well, just about ready for testing. Man this thing burns up some real estate. 😱
I need to clear off my bench before I can do any more. I also want to read back through the thread to refresh my thinking before I start testing. I want to be sure I understand how to adjust everything before I go any further. I have all three IPS sections ready but I think I will start with the VFA since it seems that is the most stable.
Blessings, Terry
Hi Terry,yours is a little different output board.
Ah yes, I see what you mean. This one is only two pair. They are the ones Jeff sent to me for testing. Hopefully the results will be useful.
Terry's building the Mini Modular output boards. It uses the same input boards as the NS series, but is a smaller simple EF3 design. It's basically the same as setting up a Slewmaster. It also can be adapted to Hexfet output devices. The input boards can be run standalone with 100R jumpers to NFB as we did with most other modular input boards.
Your tests will diffidently be useful as nobody has run this one yet. You are using BJT outputs. The LEDs on the output boards should be removed and jumper wires installed. The LEDs are for Hexfet outputs.
Your tests will diffidently be useful as nobody has run this one yet. You are using BJT outputs. The LEDs on the output boards should be removed and jumper wires installed. The LEDs are for Hexfet outputs.
Last edited:
Terry's building the Mini Modular output boards. It uses the same input boards as the NS series, but is a smaller simple EF3 design. It's basically the same as setting up a Slewmaster. It also can be adapted to Hexfet output devices. The input boards can be run standalone with 100R jumpers to NFB as we did with most other modular input boards.
Your tests will diffidently be useful as nobody has run this one yet. You are using BJT outputs. The LEDs on the output boards should be removed and jumper wires installed. The LEDs are for Hexfet outputs.
I wish I had realized this. I used MJL21193/94. Those have proven troublesome with an EF3. I will change those out before I test. These boards have been quite a journey.
I think Valery has already simulated these with MJL21193/94 and said they were fine. I need to read back through the thread to be sure though.
Are you starting to get used to surface mount parts yet?
Are you starting to get used to surface mount parts yet?
I think Valery has already simulated these with MJL21193/94 and said they were fine. I need to read back through the thread to be sure though.
Are you starting to get used to surface mount parts yet?
Yes, I have simulated MJL21193/94 with the big NS-OPS and they look fine.
Wait - I will check them in mini-ops.
I think Valery has already simulated these with MJL21193/94 and said they were fine. I need to read back through the thread to be sure though.
Are you starting to get used to surface mount parts yet?
My biggest issue with the surface mount parts is that I don't have them. Every time I think I have all the parts I need, I discover I am still missing something. I have placed orders 5 times trying to get these boards completed. I know if I want to stay in this hobby that SMD are inevitable so I am learning.
OK, my simulation shows no problem with stability, just higher 20KHz distortion, for about an order, than with 3281/1302 or Sankens.
- Home
- Amplifiers
- Solid State
- Revisiting some "old" ideas from 1970's - IPS, OPS