UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits

Notices

Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
 
Thread Tools
Old 25th Feb 2018, 4:56 pm   #1
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Grid leak bias and audio distortion.

Another thread about an audio distortion problem got me thinking about the bias method used in some radio sets in the first audio voltage amplifier after the volume control. I shall pose a question at the end of this post but first let me outline how the circuit works in terms of how I understand it.

In the first circuit below the valve is biased by a cathode resistor. The bias voltage is a function of cathode current and the value of Rk and can be calculated from ohm's law.

In the second circuit some designs save a couple of components and use grid current to bias the valve.

The third and final circuit can be used to understand how this second method works. Cg and Rg are the input components with values typically Cg = 0.01uF and Rg = 10Mohms. D [g-k] is the diode formed by the control grid and cathode of the valve.
When there is no signal input the grid and cathode are at zero volts and the cathode current is that determined by zero grid bias. What happens when a signal is applied can be understood by thinking of the circuit as a shunt rectifier circuit with Cg the reservoir capacitor, Rg the load and D [g-k] the rectifier.
The positive going part of the input signal causes the diode to conduct charging Cg. The voltage at the junction of Cg/Rg will be negative with respect to chassis. A negative voltage will be maintained here by the diode conducting on positive peaks of the input signal so long as a reasonable long time constant for Cg/Rg is chosen, hence the relatively high value (10M) of Rg.
In this second circuit the bias voltage has become a function of the input signal and the time constant Cg/Rg.

Given that this bias voltage has to be maintained by regularly driving the valve into grid current, it seems to me that it will suffer higher levels of distortion than that that might be expected from the cathode bias solution.

Now I come to my question.

What order of magnitude is this extra distortion likely to be when compared to the total harmonic distortion of 10% plus typical of the 4 valve + rectifier superhet?
Attached Thumbnails
Click image for larger version

Name:	Cathode bias.jpg
Views:	329
Size:	39.1 KB
ID:	158225   Click image for larger version

Name:	Grid leak bias.jpg
Views:	163
Size:	34.5 KB
ID:	158226   Click image for larger version

Name:	Grid leak bias equ cct.jpg
Views:	232
Size:	19.0 KB
ID:	158227  
ukcol is offline  
Old 25th Feb 2018, 5:08 pm   #2
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
Default Re: Grid leak bias and audio distortion.

What you're describing there is to me 'contact bias' - basically where the negatively-charged electrons distilled-off from the heated cathode choose to alight on the grid and thereby give it a negative charge.

It doesn't depend on signals driving the valve into grid-current, it's an intrinsic *static* characteristic of the valve.

Instances where this is used to provide bias generally have the grid-to-ground resistor of a high value - 5 to 15 Megohms - meaning electrons intercepted by the grid can't easily escape back to ground - so the grid settles at a consistently-negative-with-respect-to-the-cathode voltage.

How well it works depends on the valve's designers. Plenty of valves were designed/specified to work in this way [the 6AT6/DH77 double-diode-triode being an example].
G6Tanuki is offline  
Old 25th Feb 2018, 5:27 pm   #3
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Grid leak bias and audio distortion.

Thanks for that G6Tanuki. I'm glad I explained the circuit as I understood it as I clearly didn't understand it at all. I had already gone through all my radio and electronics text books and found an explanation of just about every biasing method but that one.

Have I understood it now if I say it simply biases the grid negative from the valves space charge with Rg high to prevent the charge leaking away too quickly?
ukcol is offline  
Old 25th Feb 2018, 6:48 pm   #4
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by ukcol View Post
Have I understood it now if I say it simply biases the grid negative from the valves space charge with Rg high to prevent the charge leaking away too quickly?
Pretty much yes - it depends on the different characteristics of the metals used for the cathode and the grid, and their relative electron-emissions at the temperatures the cathode and grid are working at.

See https://mysite.du.edu/~etuttle/elect...t27.htm#ConPot for a good [but techie] analysis.
G6Tanuki is offline  
Old 25th Feb 2018, 7:13 pm   #5
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Grid leak bias and audio distortion.

Thanks again and thanks for the link.
ukcol is offline  
Old 25th Feb 2018, 11:07 pm   #6
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Grid leak bias and audio distortion.

It is an interesting case of the same circuit doing different things depending on the design of the particular valve, the grid resistor value and the signal applied to it.

When it is a grid leak detector, normally the voltage developed across the resistor from grid current (discharging grid space charge) is relatively small and the modulated RF signal applied to the grid by the capacitor large. So the signal is rectified by the grid-cathode acting as a diode, to get the asymmetry required for AM detection, for example in a typical grid leak detector in a 1920's radio.

But if the self grid current generates enough negative grid bias (typically the grid resistor needs to be at least 3.9meg to 10 meg) and the amplitude of the drive signal is smaller, the grid stays in the negative region an no rectification or no distortion occurs for small audio signals, but it will if you overdrive the grid.

I have only few items of commercially made equipment that use this method for audio for grid bias. They did it with the EF98 in hybrid car radios (used a 10 meg grid resistor and a grounded cathode for the audio driver) it can take about 1.5V to 2V peak at the grid without rectification. Another set I have uses a 3.9M grid resistor with a 6Q7 for an audio pre-amp, the grid drive voltage is less than a volt I think.

In one other set, a very interesting trick was used to extend the signal tolerance before rectification in the grid circuit with a grounded cathode. They placed a small zinc carbon button cell in series with the grid resistor to add negative bias. Because the current is close to zero, it lasts its shelf life. They mounted it in small spring clips, soldered to a tag strip near the valve. This was done in the audio preamp stage of the 1939 pre-war TV set, the Andrea KTE-5.

Last edited by Argus25; 25th Feb 2018 at 11:14 pm. Reason: typo
Argus25 is offline  
Old 26th Feb 2018, 12:01 am   #7
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Grid leak bias and audio distortion.

Thanks Hugo.

Presumably in a half decent design the output stage will be driven into clipping before the signal is large enough to cause rectification at the control grid of the voltage amplifier stage?

Sticking my neck out again; is correct to say that the grid bias for the oscillator section of the triode/hexode mixer is obtained from rectification of the oscillator waveform at the grid?
ukcol is offline  
Old 26th Feb 2018, 1:31 am   #8
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by ukcol View Post
Sticking my neck out again; is correct to say that the grid bias for the oscillator section of the triode/hexode mixer is obtained from rectification of the oscillator waveform at the grid?
Yes that is right and why one way to test if an oscillator is likely working, is to use a high input impedance meter with a series resistor, like 47 to 100k, to measure the grid voltage; if the oscillator is working the grid voltage will read the average of the sine wave at the grid, which is substantially negative.

(One thing to watch out for using junction fets as oscillators, in circuits that resemble those for a triodes, when the gate is pushed more positive than the source, the conduction is significant causing distortion and harmonics in the output. So it is better to put a gate to source diode with a series resistor such as 470R so that the oscillator's coupling capacitor charges via that for rectification, rather than the Fet's g-s.

When Fetrons were invented to replace valves, the designers forgot about this question/issue, so if a TS6ak5 Fetron is plugged in, to replace a 6AK5 valve, it malfunctions in circuits where there is grid rectification, like a radio's local oscillator)
Argus25 is offline  
Old 26th Feb 2018, 1:38 pm   #9
daviddeakin
Hexode
 
daviddeakin's Avatar
 
Join Date: Apr 2007
Location: W Yorks, UK.
Posts: 406
Default Re: Grid leak bias and audio distortion.

With grid-leak bias you're typically operating on the curve of the grid-current characteristic, where input impedance is lower than usual, and non-linear. The distortion will therefore depend on the source impedance. It will also vary when bias shift that takes place due to the coupling cap getting charged up. The matter is further complicated by the fact that distortion taking place at the grid is complementary to that taking place within the valve itself, so you may in fact get less distortion overall, in some circumstances!
daviddeakin is offline  
Old 26th Feb 2018, 3:51 pm   #10
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Grid leak bias and audio distortion.

I seem to remember that Mullard recommended grid-current/contact potential bias over cathode bias, for one valve (EABC80? ECL82 triode? Can't remember) for this reason, for certain input levels!
kalee20 is online now  
Old 26th Feb 2018, 4:06 pm   #11
ukcol
Rest in Peace
 
ukcol's Avatar
 
Join Date: May 2003
Location: Harlaxton, Lincolnshire, UK.
Posts: 3,944
Default Re: Grid leak bias and audio distortion.

That's interesting. I would like to know the reason if anyone knows, I had always thought it was done to reduce the component count.
ukcol is offline  
Old 26th Feb 2018, 5:44 pm   #12
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by daviddeakin
The matter is further complicated by the fact that distortion taking place at the grid is complementary to that taking place within the valve itself, so you may in fact get less distortion overall, in some circumstances!
In the 1950s this idea was sometimes encouraged. The result was a reduction in second-order distortion (which we can tolerate reasonably well) and an increase in higher-order distortions (which we find most unpleasant). A poor tradeoff, but back then people didn't fully realise that not all distortions are equally unpleasant.

One reason why contact potential bias was used is that it allows the cathode to be grounded. This protects from hum due to heater-cathode leakage. Hence you will see it used in the input stage of high quality equipment, provided that it was known that the source impedance would be low enough.

Another reason was saving money. Hence you will see it used in later stages of lower quality equipment, where source impedance might not be thought about or people thought that reducing THD automatically meant better sound.
G8HQP Dave is offline  
Old 27th Feb 2018, 4:52 pm   #13
Studio263
Octode
 
Studio263's Avatar
 
Join Date: Apr 2007
Location: Hampshire, UK.
Posts: 1,574
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by G8HQP Dave View Post
or people thought that reducing THD automatically meant better sound.
Can you think of an occasion, all things being equal, where that ISN'T true? In the context of high fidelity reproduction I can't, in fact it sounds more like a cardinal goal of the whole enterprise.
Studio263 is offline  
Old 27th Feb 2018, 5:06 pm   #14
Herald1360
Dekatron
 
Herald1360's Avatar
 
Join Date: Feb 2007
Location: Leominster, Herefordshire, UK.
Posts: 16,526
Default Re: Grid leak bias and audio distortion.

I suspect that replacing 5% 2nd harmonic by 4% 3rd harmonic would not sound better, even though THD would have gone down.
__________________
....__________
....|____||__|__\_____
.=.| _---\__|__|_---_|.
.........O..Chris....O
Herald1360 is offline  
Old 27th Feb 2018, 9:11 pm   #15
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Grid leak bias and audio distortion.

Yes.

In the case of balancing anode distortion against grid current distortion, you might reduce 2nd from 2% to 0.5% at the cost of raising 4th, 5th etc. from 0.1% to 0.5%. This is likely to sound worse, although THD will have fallen.

THD is a poor measure of distortion unpleasantness, although it is not as meaningless as some 'audiophiles' claim. Various weighted versions have been proposed, but marketeers like simplicity so THD remains in use.
G8HQP Dave is offline  
Old 27th Feb 2018, 9:56 pm   #16
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Grid leak bias and audio distortion.

You described the operation of a grid leak detector properly in your first post. There is a good description here https://en.wikipedia.org/wiki/Grid-leak_detector.

The 'leak' refers to discharging of the grid coupling capacitor and not grid current from the valve which would be an extremely unreliable method of bias as it would be very valve and temperature dependent.

Your first picture is a cathode biased amplifier and would only amplify the RF signal.

The other commonly found detector in early sets is anode bend detection which uses a high value of cathode bias resistor to operate the valve at a very low anode current where it is non-linear.

The Radiotron Designer's Handbook by Langford Smith is the best source of information for valve equipment.
PJL is offline  
Old 28th Feb 2018, 6:54 am   #17
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Grid leak bias and audio distortion.

Ukol,

I think the thing to understand about a valve's bias and the issue of distortion is that for audio amplifiers, regardless of how the bias is arranged , fixed bias, cathode bias, bias from self grid current, it is important that at all times the excursions of grid to cathode signal (drive voltage) are smaller than the bias to avoid significant conduction and diode like rectification at the grid, which will 1) draw current from the driving source and 2) charge coupling capacitors and result in fairly abrupt waveform distortion.

Though some amplifiers are configured like this for class AB2 and the driver stages are more hefty to supply the grid current, but I don't think these make wonderful HiFi's though the effect is often is required to push full power out of output stages.

Still, even when the grid is run in the negative region with respect to the cathode, for a triode, with a small input signal swing, there will always be distortion in the output (plate signal).

The reason is that the dynamic transfer characteristic of the valve (grid voltage versus plate current, with plate voltage held constant) is not a straight line, but curved, more, like a parabola. So if the grid is driven with a symmetrical sine wave, the negative going part of the drive waveform produces less of a change in plate current than the positive going half.

Because of the more or less parabolic shape of the transfer characteristic it turns out that the non linear distortion contains a component of twice the signal frequency and in addition, due to the asymmetry, the applied sinusoidal signal alters the average value of the plate current and some rectification takes place. Also if two frequencies are present in the drive voltage , sum and difference components occur or intermodulation frequencies. (the rectification effect is exploited in the anode bend detector)

So considering all that, it is a wonder that most class A triode amplifiers sound reasonable at all, but they do, as the percentage of second harmonic distortion is mostly small enough to be tolerable.

In amplifiers with large signal swings, higher order odd harmonics are generated and Fourier analysis of these shows them to contain odd sine components and even cosine components when the excitation is a sine wave.

Fortunately, for audio signal amplification with valves, we are not stuck with the non-linearity of triodes. A pentode can be configured for zero 2nd harmonic distortion by the selection of the correct load resistance and operating conditions. Interesting that this is impossible with a current excitation device like a transistor. I have mentioned this fact before on other threads.

Last edited by Argus25; 28th Feb 2018 at 7:01 am.
Argus25 is offline  
Old 28th Feb 2018, 10:55 am   #18
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by Argus25 View Post
So considering all that, it is a wonder that most class A triode amplifiers sound reasonable at all, but they do...
Yes it's amazing to me, too.

Push-pull, and using grid distortion to cancel Ia/Vg distortion, can indeed reduce 2nd harmonic distortion to zero, but other distortions remain, and their presence can be more objectionable than 2nd harmonic distortion ONLY, even if THD has gone down. As Herald1360 has already observed!

Distortion isn't a simple matter. Negative feedback can increase some types of distortion even though THD can go down, so where the figures say one thing, our ears tell us another.

I have a Grundig 2043 which sounds amazing, though the EABC80 uses contact potential bias! The EABC cathode needs to be tied direct to chassis, because of the detector/discriminator circuitry, so cathode bias is a no-no. But although back bias could have been used, the Grundig designers just went for a 10M resistor.
kalee20 is online now  
Old 28th Feb 2018, 11:05 am   #19
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Grid leak bias and audio distortion.

As far as I know all capacitor coupled cathode biased PP audio amplifiers can be driven into rectification on the grid. I think the Mullard 5-20 documents describes the consequence of this which is to reduce the bias and the overall dissipation in the output valves. This is not the same as RF amplifiers where the driver stage is designed to push the valve into grid current to operate the valve in this region of the characteristics.

Standard distortion tests with constant amplitude signals will not detect the consequences of grid rectification. I am sure there are articles out there somewhere!
PJL is offline  
Old 28th Feb 2018, 12:14 pm   #20
Argus25
No Longer a Member
 
Join Date: Oct 2016
Location: Maroochydore, Queensland, Australia.
Posts: 2,679
Default Re: Grid leak bias and audio distortion.

Quote:
Originally Posted by PJL View Post
As far as I know all capacitor coupled cathode biased PP audio amplifiers can be driven into rectification on the grid. I think the Mullard 5-20 documents describes the consequence of this which is to reduce the bias and the overall dissipation in the output valves.
When the signal is AC coupled via capacitors to the grid, then grid rectification on audio peaks pushes the bias, on the average, more negative depending on the audio content. In an odd sense this is like a type of AGC which tends to power down the output valves and as you say, reduce dissipation.

One of the age old problems in audio work of course is the enormous difference between music as a signal, vs a test laboratory sine wave of uniform amplitude. With music there is an enormous dynamic range and the peaks can be subject to either clipping and/or rectification, while the average power level is still below a sine wave that would not as yet induce clipping. This is partly why laboratory tests on audio amplifiers might sometimes disagree with listening tests.

(I could have added to my post above, one of the wonderful things about the nearly parabolic or y = x^2 relation relation between grid voltage and plate current of a triode, is that when two different frequency signals are presented to the grid, the plate current contains the sum & difference frequencies, which makes it possible to have a single triode as a mixer in a radio or TV. The same applies to Fets & transistors. So what we think of as "distortion" in the audio sense, is hugely beneficial in radio work and results in simple and elegant circuit configurations, just using one active device).

Last edited by Argus25; 28th Feb 2018 at 12:30 pm.
Argus25 is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 4:27 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.