I intend to build an MC phono stage, in LTP topology, a-la the late Allen Wright's RTP3C. That is, LTP JFET's, cascoded by tubes, with some changes.
It's going to be transformer loaded and in Ultrapath configuration, a-la Lynn Olsons's Raven PreAmp.
I thought about using a pair of triode-strapped C3g tubes per channel, cascoding a pair of JFET's.
D3a will probably have too much gain for this configuration. Also, I read reports that they are a bit microphonic, which I'd rather avoid.
Did anyone measure those tubes for linearity?
I'm not concerned with the amount of the lower harmonics distortion percentage, I'm concerned with the amount of higher harmonics distortion percentage.
All information based on actual experience, especially in phono stages, will be highly appreciated.
It's going to be transformer loaded and in Ultrapath configuration, a-la Lynn Olsons's Raven PreAmp.
I thought about using a pair of triode-strapped C3g tubes per channel, cascoding a pair of JFET's.
D3a will probably have too much gain for this configuration. Also, I read reports that they are a bit microphonic, which I'd rather avoid.
Did anyone measure those tubes for linearity?
I'm not concerned with the amount of the lower harmonics distortion percentage, I'm concerned with the amount of higher harmonics distortion percentage.
All information based on actual experience, especially in phono stages, will be highly appreciated.
In a phono stage with typical smallish signal levels mathematics comes to your rescue and more or less guarantees low amounts of high order distortion unless you get the biassing badly wrong.
It is NOT a simulation. These are real C3g curves from Ale`s excellent curve tracer.
You have everything and calculate what you need.
Thank you.
Indeed, with small signal levels, distortion is low. However, various tubes have different distortion pattern, which is why I'd like to see some measurements and/or listening impressions in comparison to other tubes in phono stage.
Thank you.
I have no idea how to figure out the distortion, especially the distortion amount of various harmonics, from the anode curves.
Does anyone have any idea how to do it?
Much has on the net. Here's the first from Google search.
You have for triode also.
Of Loadlines Power Output and Distortion - Part 5
As a general rule you can't get higher order distortions from load lines etc. as they are only approximations for the purposes of biasing, and subject to wide sample variation. An estimate of 2nd and 3rd is about all you can do. Beyond that you just have to measure individual samples.
Your posts are unoptimistic as usual.
Good charts with the help of a PC are sufficient for the computation of high order harmonics.
Some people are optimistic, others are pessimistic, while some are realistic.
1. How do you obtain good charts?
2. What PC software are you referring to?
3. Have you tried it and compared it to actual measurements?
Hi Joshua_G,
Real measurements two dozens of used C3g tubes (C3g triode connected, B+: 300V, 11mA CCS anode load, 235R and 100uF in the cathode):
Uin: 15mV RMS THD: typical 0.05-0.12%
Uin: 100mV RMS THD: typical 0.028-0.068%
Uin: 1V RMS THD: typical 0.14-0.28%
Real measurements two dozens of used C3g tubes (C3g triode connected, B+: 300V, 11mA CCS anode load, 235R and 100uF in the cathode):
Uin: 15mV RMS THD: typical 0.05-0.12%
Uin: 100mV RMS THD: typical 0.028-0.068%
Uin: 1V RMS THD: typical 0.14-0.28%
Hi,
Thank you very much.
It looks very good.
1. How the THD figures compare to other triodes, or triode-strapped pentodes?
2. Do you have the harmonic distortion spread over various harmonics?
3. Did you listen to them? If yes, in what kind of amp and how did they sound, both by themselves and in comparison to other tubes?
Thank you very much.
It looks very good.
1. How the THD figures compare to other triodes, or triode-strapped pentodes?
2. Do you have the harmonic distortion spread over various harmonics?
3. Did you listen to them? If yes, in what kind of amp and how did they sound, both by themselves and in comparison to other tubes?
Nonetheless, he's correct. Since one can see very different distortion amounts and distributions among samples of the same nominal type of tube (see, for example, the survey of 6SN7 results in "Valve Amplifiers", where there's over 10dB difference in 3rd and 4th order distortion), the idea that one can make these predictions from models is not borne out by experiment.
It's even more difficult for small signal swings, where just about any tube gives vanishingly low distortion. The first stage tube in a phono stage needs to be chosen for low noise, low microphonics, and high gain, not large signal distortion.
I try to be realistic. You can extract numbers from a chart but they are likely to include no significant figures for higher orders, as I said. The fact that software can provide numbers does not mean that they are useful. Given a particular valve sample the best method is to simply measure the output with an FFT. Carefully measuring the valve curves and then calculating distortion from the curves is possible but much more work. Another sample will gve different results. Curves from a simulator will not tell you much.RajkoM said:
The OP understands this, hence his question about actual measurements.
We're talking about real C3g curves from Ale Moglias curve tracer, not simulated curves. Unless you think that Ale simulated curves ???
Some changes between C3g tubes are possible, ie. NOS Telefunken triple mica C3g have substantially smaller Ri than newer Siemens two mica C3g........etc.
End of story from my side.
Real curves either apply to one valve - the one they were measured from - or no valves (if they are an average of several samples). In neither case do they apply to the valves which the OP is using.
Given a set of measured points one can always find a curve fit to them, given a sufficient order. The higher order coefficients, from which a computer can calculate higher order distortion, contain almost no significant figures as they are essentially fitting noise caused by small measurement errors. This is elementary experimental science.
Given a set of measured points one can always find a curve fit to them, given a sufficient order. The higher order coefficients, from which a computer can calculate higher order distortion, contain almost no significant figures as they are essentially fitting noise caused by small measurement errors. This is elementary experimental science.
- Status
- Not open for further replies.