Theoretical question

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
abbey road d enfer said:
Your voltmeter is probably rectifying. Does the oscilloscopê show this DC shift?


Well, I am sorry to say that I did not measure this VDC with the scope just only with the voltmeter.

As you suggest that the voltmeter "can" rectify I surely would like to do a second check sometime, but that will include changing cables again :)

But even if it was rectifying… wouldn't 4VDC be a little on the low side at full gain?


Best regards

/John


Best regards

/John
 
If voltmeter rectification was eliminated then I would speculate on the clipping of the output transformer. This will raise the  right hand side of the 2.2uF to a DC potential.

In situations like this without a scope is like shooting in the dark.
 
johnheath said:
But even if it was rectifying… wouldn't 4VDC be a little on the low side at full gain?
I was referring to rectification due to non-linear behaviour of the input stage of teh voltmeter (assuming it was a digital VM); there is almost no correlation between the HF voltage and the apparent DC value.
 
Thanks both of you

I guess there are a couple of reactions on the fact that the signal was interfered right after the input transformer. The problem was solved by the use of the shielded cable and I have more information now on how to use the scope even more and not rely on voltages measured by the volt meter.

It has been a great theoretical study for me… thanks again.


Best regards

/John
 
You can't really have DC on the grounded OT primary. This suggests your meter is confused by something. Many "DC" meters are confused by AC much larger than DC.

The main issue is that this is a HIGH gain circuit. About 2,000 open-loop, about 300 closed-loop. The least sneakage from the output plate-leads back to the first grid begs for oscillation.

When a tube oscillates, the grid can not go much positive because it conducts as a diode. But it will swing way negative. The average DC shifts.

Layout is the main thing. Don't run your sewer near your drinking water, keep big-end outputs far-far from small-end inputs. A classic problem is assuming that OT primary leads may run all over. Sure they do not mind, but they radiate back into the input.

The input section should be exceptionally compact and tight. I would butt the input transformer next to the tube, short-short wires. However you have a lot of "stuff" here.

I do NOT see what either "220nF" cap does for you. No DC isolation is needed either place, and the bass-cut is far below the audio band.
 
PRR said:
You can't really have DC on the grounded OT primary. This suggests your meter is confused by something. Many "DC" meters are confused by AC much larger than DC.

The main issue is that this is a HIGH gain circuit. About 2,000 open-loop, about 300 closed-loop. The least sneakage from the output plate-leads back to the first grid begs for oscillation.

When a tube oscillates, the grid can not go much positive because it conducts as a diode. But it will swing way negative. The average DC shifts.

Layout is the main thing. Don't run your sewer near your drinking water, keep big-end outputs far-far from small-end inputs. A classic problem is assuming that OT primary leads may run all over. Sure they do not mind, but they radiate back into the input.

The input section should be exceptionally compact and tight. I would butt the input transformer next to the tube, short-short wires. However you have a lot of "stuff" here.

I do NOT see what either "220nF" cap does for you. No DC isolation is needed either place, and the bass-cut is far below the audio band.

Well, I do not know if my rather expensive Keysight volt meter is a crappy one or not but as I said I have learned today not to trust it more than the scope… appreciated information really.

Well, I also learned more about oscillation and have experienced that many times before and know some trick to avoid it.

The gain of this circuit, here I might be wrong, is in fact measured to 60dB (1000?) or I am doing something completely wrong in calculating or measuring or hooking up the probes from the scope. The sneak FB is actually measured to get the gain down to just 60dB since that is quite enough for my needs. A 390k resistor brings the measured gain down to 56dB. A smaller value would probably bring down the gain even more of course but then it would be too low perhaps? The measuring is in my case including the input and output transformers.

The input transformer is not "butted" so close to a tube but the input circuit is compact in its layout with very short leads and from that I usually use shielded cables… also to avoid interference from the output transformer which sits rather close to the  input transformer… both as far away as possible from the power transformer and PSU.

I would not argue about the two 220nF caps because as you say they are not needed to block any DC… it should have been taken care of earlier. in the circuit, but the bass cut is quite audioable to my ears… or am I missing something here? 1/(220k*6,8nF*pi*2) = 106Hz and 1/(220k*13,6nF*pi*2) = 53Hz? I can both hear it and see it on the scope that it falls 6dB when switched on.

I hope that I am not on too thin ice here.

Best regards

/John



 
johnheath said:
Well, I do not know if my rather expensive Keysight volt meter is a crappy one
Only the very best VM's behave when presented with a high does of RF.
or not but as I said I have learned today not to trust it more than the scope
I would say most of the times, the scope is more reliable (but less accurate), because its operating bandwidth is higher than that of a DVM.

The gain of this circuit, here I might be wrong, is in fact measured to 60dB (1000?)
And this is a lot; parasitic coupling becomes a serious concern. It takes only a few picofarads of stray capacitance to start a feedback oscillation.


I would not argue about the two 220nF caps because as you say they are not needed to block any DC… it should have been taken care of earlier. in the circuit, but the bass cut is quite audioable to my ears… or am I missing something here? 1/(220k*6,8nF*pi*2) = 106Hz and 1/(220k*13,6nF*pi*2) = 53Hz? I can both hear it and see it on the scope that it falls 6dB when switched on.
I believe PRR was referring to the other HPF created by the 220nF cap and the 1 Meg resistor that follows. They define a cut-off at about 0.7Hz, which probably, although being harmless, doesn't seem useful either. You could ditch the 220nF caps, particularly the one at the input, since it adds stray capacitance to the already very sensitive input node.
 
"I believe PRR was referring to the other HPF created by the 220nF cap and the 1 Meg resistor that follows. They define a cut-off at about 0.7Hz, which probably, although being harmless, doesn't seem useful either. You could ditch the 220nF caps, particularly the one at the input, since it adds stray capacitance to the already very sensitive input node."

Oh, yes that is also some sort of bass cut… way below the audio band. Those two 220nF coupling caps has a quite newbie-stamp on there and have a slightly embarrassing history in fact. I might as well come clean with it… it might be a lesson for a person being in the same situation as I was when they were "made".

It is from a time a couple of years ago when I was trying to figure out a HPF and tried to place it around the gain pot between two gain stages… at the same time as using a FB between the same gain stages… did not know then that it is far more tricky to calculate and set up as without the FB. Anyway… what I found out was that when measuring the resistance of a pot or shunt resistor in parallel with another the resistance was lower than the lowest resistance (as it should be) but when I placed a cap in between the two resistances they could be measured one at the time (with the volt meter that is).

For calculation reasons they stayed in the circuit… I know better these days, but I had to use a cap before the second resistor, large enough not to affect the bass response… which makes a useless bass cut ;)

The circuit on that schematic is a couple of years old and has stayed because it sounds great and actually I have always used shielded cables from the input to the grids and really had no problem with it even though it has plenty of gain as you say.

I am a guy trying learn this all by my self  (too old for school again) and sometimes that means taking the heat from asking stupid questions and revealing lack of knowledge and all… but I can live with it :)

But , just another related question. If 60dB of gain is plenty… How come that there are so many preamps with gain in the region of 72dB - 80dB?… that seems a LOT to me.


Best regards

/John

 
johnheath said:
But , just another related question. If 60dB of gain is plenty… How come that there are so many preamps with gain in the region of 72dB - 80dB?… that seems a LOT to me.
Agreed; how much is a lot? The right answer is "it depends", as always.
People doing nature recording want to have a lot of gain, in order to capture minute sounds. They often have to use dynamic mics, for their robustness, so they need a lot of gain. With 80 dB gain , the dynamic range cannot excess about 48 dB; taking into account some necessary headroom, it results in about 40dB effective S/N ratio - that's about the S/N ratio of a cassette recorder without Dolby. So that's about the limit for acceptable results.
For classical music, recording with distant mics, 60 dB gain is about the maximum, resulting in 60dB effective S/N ratio in the soft sections. Again, it seems to be the acceptable limit considering the typically high expectations of classical music lover.
For rock/jazz/blues/pop/country, the use of close_micing significantly changes the gain requirements. I would say that in 99% of the cases, the gain is comprized between 20 and 50 dB. In fact I reckon I could use a 35 dB fixed-gain mic pre 9 times out of 10. That is, indeed, using digital recording. If I had to print to tape, I would really need the 30 dB range of adjustment.
 
abbey road d enfer said:
Agreed; how much is a lot? The right answer is "it depends", as always.
People doing nature recording want to have a lot of gain, in order to capture minute sounds. They often have to use dynamic mics, for their robustness, so they need a lot of gain. With 80 dB gain , the dynamic range cannot excess about 48 dB; taking into account some necessary headroom, it results in about 40dB effective S/N ratio - that's about the S/N ratio of a cassette recorder without Dolby. So that's about the limit for acceptable results.
For classical music, recording with distant mics, 60 dB gain is about the maximum, resulting in 60dB effective S/N ratio in the soft sections. Again, it seems to be the acceptable limit considering the typically high expectations of classical music lover.
For rock/jazz/blues/pop/country, the use of close_micing significantly changes the gain requirements. I would say that in 99% of the cases, the gain is comprized between 20 and 50 dB. In fact I reckon I could use a 35 dB fixed-gain mic pre 9 times out of 10. That is, indeed, using digital recording. If I had to print to tape, I would really need the 30 dB range of adjustment.

That is very interesting information. I am not a recording expert what so ever but the jobs I have done have seldom required very much gain from the preamp. I mostly record acoustic music and vocal.

Good info… thanks

Best regards

/John
 
johnheath said:
...I am a guy trying learn this all by my self  (too old for school again)...............

/John

Never too old.

I am doing a mathematics module at open university and will be going back to the uni in September to do EEE.
 
sahib said:
Never too old.

I am doing a mathematics module at open university and will be going back to the uni in September to do EEE.

Well, true in a sense... but I guess that a EEE would mean full time studies? In sweden you can use the studying welfare up to 44 years old but I am 46 and  I earn good money as I am as a police officer. It was hard enough to study at the police academy alone with a child living like a poor man. And do that again???...Noooo, not really :)

But you are true that you are never too old to learn though.


Best regards

/John
 

Latest posts

Back
Top