(How) could I get a balanced power amplifier to oscillate as a "black box" somewhere around 1 kHz for simulation purpose ?
I want to use only + in and - in, + out and - out. No use of ground. Not touch the amplifiers feedback if possible.
I want to use only + in and - in, + out and - out. No use of ground. Not touch the amplifiers feedback if possible.
I proposed that long ago 🙂
However I‘m so busy with something else that I could not work on that.
Was hoping someone else does…
I think I posted schematics of post 1 kHz low pass filters…
I'm intrigued by your question.
First thought was a Victor oscillator, powered from batteries to avoid grounding issues, but I'm betting that somehow misses the objective.
Would you elaborate a bit? I'm guessing the amp in question isn't a generic hifi power amp?
It is a hifi amplifier with balanced inputs and balanced outputs.
My ideas behind that:
1. If simulation of a oscillator based on said amplifier shows very good results, it is a sign that the amplifier must be very good too.
2. If this is the case, a scaled down version could be used to built a low distortion oscillator.
I think the logic is flawed.
A power amp will always have much less performance, more distortion and noise than a dedicated low distortion/noise oscillator.
The topologies and requirements are completely different.
Starting with a power amp and scaling down will not get you a low distortion/noise oscillator like the Viktor.
Jan
A power amp will always have much less performance, more distortion and noise than a dedicated low distortion/noise oscillator.
The topologies and requirements are completely different.
Starting with a power amp and scaling down will not get you a low distortion/noise oscillator like the Viktor.
Jan
Because: 1. If simulation of a oscillator based on said amplifier shows very good results, it is a sign that the amplifier must be very good too.
Noise can be removed by averaging, possible higher voltage swing would give lower noise ?
Still, I believe in the passive post filter.
Maybe because of the high gain of the opamp used by Victor.
"scaled down" means it is not a power amplifier anymore, the necessary gain could be set.
The problem is that simulation always assumes perfectly matched components, and no PCB layout effects. Also it will also greatly depend on how sensitive the design is to component variations. Victor saw that when designing his oscillators.
Then also, the op amp open loop gain is both extremely high and has enough phase angle margin to be stable at the gains used. The extra gain goes directly to lowering distortion. Today's op amps would be next to impossible to beat for noise and gain. Distortion too, looking at even a decade old op amp design.
To be honest, the PCB layout and power supply will impact thing more.
Then also, the op amp open loop gain is both extremely high and has enough phase angle margin to be stable at the gains used. The extra gain goes directly to lowering distortion. Today's op amps would be next to impossible to beat for noise and gain. Distortion too, looking at even a decade old op amp design.
To be honest, the PCB layout and power supply will impact thing more.
Hello,
Has anyone heard from Victor recently?
I sent him an email at the beginning of August about the oscillator and he replied. Then no more news from him to complete the transaction. I restarted him twice with no response.
I checked my spam folder just in case. Nothing.
Best Regards,
Stef.
Has anyone heard from Victor recently?
I sent him an email at the beginning of August about the oscillator and he replied. Then no more news from him to complete the transaction. I restarted him twice with no response.
I checked my spam folder just in case. Nothing.
Best Regards,
Stef.
Hi ... just in case Viktor happens to not see your post:
https://viccc42.wixsite.com/uld-audio
Cheers, Jesper
https://viccc42.wixsite.com/uld-audio
Cheers, Jesper
gentlevoice Thanks, I can't seem to give your post a thumbs up.
I have sent Victor an email from your link above.
I have sent Victor an email from your link above.
I figure this thread is about as good as any to post a recent observation.
My balanced out ULDO oscillator with "Nacho" notch filter THD typically measures around -139 dBc (0 dBc=+20dBu) driving 100 mW into 600Ω and about -141 dBc into 10KΩ. It's primary use is in pro audio where balanced output is a must-have. It's not in the -150 dBc range of Viktor's but still a good 20-30 dB better than most gear it measures. Most of the harmonic content is HD3.
I've been using Yageo MFR-25 resistors with great results and hand-select two values to reduce a trim range and provide better frequency accuracy. The two values are in the bridge arms (15K+909R and 15K+909||15K+909) and two 10K in the inverter stage. Eight resistors total.
I recently decided to try 0.1% Dale CMF55-series to avoid hand-selection and found the THD measured at least 15 dB worse with the Dales. The added distortion was all HD3.
These two resistor locations - as I'm sure Viktor can also attest to - are the most critical for low distortion.
You read here that resistor distortion is related to power - and this true - but the maximum dissipation in the 10KΩ feedback resistor is only about 1.5 mW "dissipated" in a 500 mW resistor. That's "25 dB of power headroom" LOL. The dissipation in the two bridge arms even less.
I would have captured an FFT but to be honest was so anxious to get the Dales out that I didn't. Sure enough replacing them with the Yageo brought it back down into the typical range.
So the question becomes what's a better resistor in 0.1% 250 mW through-hole? Holsworthy? https://www.mouser.com/ProductDetail/279-YR1B10KCC
Not going to do the bulk metal foil thing as they're crazy-expensive.
My balanced out ULDO oscillator with "Nacho" notch filter THD typically measures around -139 dBc (0 dBc=+20dBu) driving 100 mW into 600Ω and about -141 dBc into 10KΩ. It's primary use is in pro audio where balanced output is a must-have. It's not in the -150 dBc range of Viktor's but still a good 20-30 dB better than most gear it measures. Most of the harmonic content is HD3.
I've been using Yageo MFR-25 resistors with great results and hand-select two values to reduce a trim range and provide better frequency accuracy. The two values are in the bridge arms (15K+909R and 15K+909||15K+909) and two 10K in the inverter stage. Eight resistors total.
I recently decided to try 0.1% Dale CMF55-series to avoid hand-selection and found the THD measured at least 15 dB worse with the Dales. The added distortion was all HD3.
These two resistor locations - as I'm sure Viktor can also attest to - are the most critical for low distortion.
You read here that resistor distortion is related to power - and this true - but the maximum dissipation in the 10KΩ feedback resistor is only about 1.5 mW "dissipated" in a 500 mW resistor. That's "25 dB of power headroom" LOL. The dissipation in the two bridge arms even less.
I would have captured an FFT but to be honest was so anxious to get the Dales out that I didn't. Sure enough replacing them with the Yageo brought it back down into the typical range.
So the question becomes what's a better resistor in 0.1% 250 mW through-hole? Holsworthy? https://www.mouser.com/ProductDetail/279-YR1B10KCC
Not going to do the bulk metal foil thing as they're crazy-expensive.
Yageo MFR-25 come in different TempCo, versions I usually get tthe 50ppm ones since they are so cheap.
Its so ironic that the cheapest resistor gives better performance. Many folks hold those CMF55 as the gold standard.
based on what you have measured I'd say it's a crap shoot, each has to be tried and measured to be sure.
How about Vishay MBB0207 and MELF MMB0207?
Its so ironic that the cheapest resistor gives better performance. Many folks hold those CMF55 as the gold standard.
based on what you have measured I'd say it's a crap shoot, each has to be tried and measured to be sure.
How about Vishay MBB0207 and MELF MMB0207?