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 20th Jul 2018, 4:51 pm   #21
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cathode resistors and bypass caps.

The blocking cap won't cause you any problem Andy, it's the capacitance to ground (strictly speaking to AC ground) in parallel with the load that bites when you use an ECC83. The high ra is the price you pay for the high mu (100) and the low gm (1.6mA/V).

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 20th Jul 2018, 6:41 pm   #22
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by Diabolical Artificer View Post
L, said decoupling cap is de-coupling/blocking DC from test lead, not in series with the anode, as you probably knew/guessed.
Yes, I figured as much, thanks for the clarification.

Lawrence.
ms660 is offline  
Old 20th Jul 2018, 8:10 pm   #23
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Cathode resistors and bypass caps.

Show us the complete circuit, including sources and loads and bias conditions. Then we can calculate the expected gain and frequency response.
G8HQP Dave is offline  
Old 20th Jul 2018, 10:55 pm   #24
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,059
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by G8HQP Dave View Post
Yes, for pentodes it gets complicated. Scaling 1/gm by the current ratios seems about right, provide that the screen is decoupled to ground. If decoupled to the cathode then there is some positive feedback so I'm not sure what happens then.
No, it would be negative feedback! For normal pentode operation, screen should be decoupled to the CATHODE, so that cathode-screen potential stays constant, and you can use the appropriate valve characteristic curves.

If you decouple screen to GROUND, then applying 1V rise to the control grid, the cathode voltage rises (by something a bit less than 1V). But if the screen is held constant relative to ground, then the cathode-screen voltage will then have fallen a bit, which acts to reduce valve current marginally and thus cathode potential, not raise it.

In fact, if you decouple screen to ground, in a cathode follower, you are approaching triode operation! Since anode voltage is steady, and now you have made screen voltage steady by bypassing to ground, then from an AC point of view, anode and screen are at the same potential. You just might have them at different DC levels.

Last edited by kalee20; 20th Jul 2018 at 10:56 pm. Reason: clarity
kalee20 is online now  
Old 21st Jul 2018, 12:19 pm   #25
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Cathode resistors and bypass caps.

"The blocking cap won't cause you any problem Andy, it's the capacitance to ground (strictly speaking to AC ground) in parallel with the load" That'll be the internal capacitance multiplied by gain yes?

"Show us the complete circuit, including sources and loads and bias conditions." Will do Dave.

Pentodes look a nightmare Paul, I'm struggling a bit with triodes so will tackle pentodes another time but get the reference to the original question.

I've tidied up my test setup making connections shorter, especially from the OP to the AC voltmeter, HF attenuation as a result is a lot better. Reducing the de-coupling cap to 10n improves things again.

In an effort to get a handle on the circuit better I had a go at calculating the various parameter's. The problem is calculating Ra which you need to further calculate OP Z etc. In Langford Smith's Radio Designers and other sources Ra is estimated by drawing a line ( from the load line) following the grid curves to get delta Va/delta Ia when the bias voltage is constant. I found this really tricky, it's guesswork and any error throws out the rest of your calculations. Is there a better way?

I kept at it and eventually got an Ra of 59k with a gain of 32 with Rk unbypassed, I also managed to calculate Cin = 60p (52p + stray) and Zout 210k, which seems high. With Rk at 1.5k Vgk should be 1.5v and Va 190v in reality it's 1.3v and 210v. With a gain of 32v OP V should by 17v RMS for an IP V of 0.556v, I'm getting 14v RMS. So therefore there is some disparity, but that's valves to a point.

To cut to the chase, to better calculate the following stages grid resistor and coupling cap to a following stage, if needed and to work out the frequency response etc I have to do some more calculations, however I guess all this changes with NFB applied, which I'll put to one side for now. Any tips welcome, how do others go about designing a valve IP stage/voltage amp?

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 21st Jul 2018, 1:04 pm   #26
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

For ra calcs from valve characteristics, if you're down on the bend then tangent line, if up on the "straight" part and it's relatively straight then no line needed for ra calcs provided Va doesn't drop to that part of the curve that isn't as straight.

Lawrence.

Last edited by ms660; 21st Jul 2018 at 1:10 pm.
ms660 is offline  
Old 21st Jul 2018, 1:28 pm   #27
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by Diabolical Artificer View Post
... Any tips welcome, how do others go about designing a valve IP stage/voltage amp?
You have to start by deciding how much gain you need in that stage, what frequency range you'd like to cover and what noise and distortion levels you can accept. There might also be some power supply limitations - for example if you really don't want to run with 10mA quiescent current in the valve then you won't be able to make the best of an alternative like the ECC81 (this has nearly as much gain as the ECC83 - a mu of 70 rather than 100 - but a significantly higher gm so ra is correspondingly lower).

All that said, it is possible to use the relatively weedy ECC83 in a 'heavy lifting' role. The Velleman K4000 and K4040 amps use just one as the driver for four EL34s delivering 90W in parallel/push-pull ! You just have to minimise the load capacitance and use plenty of global feedback to sort the frequency response out.

EDIT: It's also worth saying that better (and many, many more) men than me worked their way through this problem back in the day. You could do worse than look up the solutions that they found - Harold Leak did all the small-signal work in his Stereo 20 using only ECC83s, for example.

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 21st Jul 2018, 1:51 pm   #28
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,951
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by ionburn View Post
I think the cathode bypass cap is often not used on output stages for the reasons you indicate..
I always thought the reason for not using a cathode-resistor bypass-cap in an output stage in a radio/record-player was that when you drive the valve hard (AB1) for maximum output, the _average_ current through the cathode resistor ceases to be constant, charging the capacitor and causing disruption to the bias characteristics for a period which depends on the R/C timeconstant.

Unbypassed cathode-resistors don't suffer from this effect because there's no capacitor to give a R/C timeconstant!

[It's the same as why - in classic transformer-coupled push-pull transistor amps - the sort you get in 1960s transistor-radios with a pair of OC72s, OC81s or whatever, the emitter-resistor that's common to both output transistors is never bypassed even though it potentially dissipates a significant percentage of the audio power].
G6Tanuki is offline  
Old 21st Jul 2018, 5:55 pm   #29
ionburn
Heptode
 
ionburn's Avatar
 
Join Date: Dec 2014
Location: Lincoln, Lincolnshire, UK.
Posts: 583
Default Re: Cathode resistors and bypass caps.

Yes you may well be right. I was just aware of a reduction in possible distortion by not having the cap, but am not aquainted with the details.
ionburn is offline  
Old 21st Jul 2018, 6:22 pm   #30
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

Andy, if not already seen, the RCA tube manual might be of interest for a bog standard voltage amplifier stage, the link below gives quick access to RC30, details start at book page 641, the component values and voltage data for the valves covered are a bit further on in the book, you can't miss them:

https://archive.org/stream/RCA_RC-30...1975#page/n641

Lawrence.

Last edited by ms660; 21st Jul 2018 at 6:29 pm.
ms660 is offline  
Old 21st Jul 2018, 8:52 pm   #31
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: Cathode resistors and bypass caps.

There is another effect working towards roll-off at higher frequencies, Andy.

The source generator and the circuitry between it and the grid will have an impedance. There is the obvious grid to cathode capacitance and the strays of the wiring, but the grid to anode capacitance is more insidious. It isn't simply capacitance to ground, the voltage of the anode is zooming off in the opposite direction, and amplified by the gain! The result is like having a much bigger capacitor to ground... much bigger by a factor equal to the gain. So a relatively small Cga and the strays in wiring can add a surprising amount of roll-off. This is called the Miller Effect.

With an ECC83, and a high anode resistor and a high load impedance, the roll-off point can be disappointingly low. To get wide bandwidth takes serious work.

Making wideband amplifiers is serious work. Un-decoupled cathode resistors are a form of feedback and flatten the gain (while reducing it). So wideband systems have a lot more stages, with less gain per stage. The number of stages means that the flatness of each stage needs to be better and that means more degeneration and less gain and more stages.

Degeneration as all other forms of feedback reduces gain, flattens gain with frequency, reuces distortion, and does nice things to input and output impedances. It's just that this form of feedback isn't made with a visible loop path, it's done by having common impedance in both the input and the output paths.

It makes a mockery of most of those people vigorously opposed to feedback. It's all around them and most of them fail to see it. Show them a circuit with a path back to an earlier stage and their crosses and cloves of garlic come out. Electronic design is a science. There is enough freedom of choice that there's also a good bit of art involved. You just have to dodge the attempts of people who can't do the science or the art, but try to make it a mythology and a religion.

Becoming interested in the effect of un-decoupled cathode (or emitter) resistors is a big step of understanding.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 22nd Jul 2018, 8:01 am   #32
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Cathode resistors and bypass caps.

This drawing a line sort of following the grid line on a mutual charactoristics graph to find ra drives me nuts Lawrence, and what do I use as my fixed bias, actual or as on the load line?. It's arbitery guesswork, how far you extend it vertically obviously effects the calculation, and how far you extend it is pure whim ... ish. Can I measure it though? How about I change the HT whilst keeping the bias constant then read off V and I ? I can do this as I'm using a HV PSU.

Thanks G, will do.

Re post # 30 I did this for the front end Lawrence in my first amp using values off the Philip's datasheet which is extensive for the ECC83, I'll check the RCA one. This worked to an extent, I can knock up a valve amp using a cookbook like the Valvewizard as easy as falling off a log, stick music in one end, it comes out the other, but I wasn't happy with the frequency response, though distortion/noise/crosstalk etc was/is excellent.

C in I calculated as 60 puff David, which surely only attenuates the signal at quite high frequencies, haven't checked yet though.

"Making wideband amplifiers is serious work" I'm finding this, but would you class 20hz ish to 20khz wideband? To be honest am not sure why amp designers bother to go this high, most of us can't hear that high as you know, only youngsters who are all listening to crap pop music anyway and couldn't give a stuff about fidelity.

So your saying a bit iof degeneration is a good thing? Degeneration I presume is loss or attenuation?

"Becoming interested in the effect of un-decoupled cathode (or emitter) resistors is a big step of understanding." It is. This all started with a cascode which was part of an intended cascode LTP a la Hedges but I thought it better to try and suss out an easier config like the common cathode gain stage using the ECC83 which has very good data first.

I'll do some more work/testing today, this is taking me down the testing/testgear sideroad as well, interesting stuff.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 22nd Jul 2018, 10:34 am   #33
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

Andy, re: ra calcs from manufacturers valve data, didn't have a print out of the ECC83 characteristics to hand but I had one for a 6Q7, valve data figures gives ra as 58k with Va at 250v and Vgk at -ve 3v, at those voltages Ia is given as 1mA....first attempt at a tangent line to the Ia/Va curve for that bias at 250v intersected Va at 195v, the operating point on the curve also corresponded to an Ia of 1mA as given in the valve data figures, so that's a 1mA change of Ia for a Va change of 55v = 55k for ra.

Next I did the calc without drawing a tangent to the curve, I used a change of Va of 20v (10v either side of 250v) that returned a change in Ia of 0.36mA = 55.5k for ra, those two results are close enough to the published figure considering the thickness of the lines on the print out and the scale and possible scale distortion due to original scan/upload/download/printer/wobbly hands etc.

If you print out the ECC83 characteristic curves you should be able to do similar.

Lawrence.

Last edited by ms660; 22nd Jul 2018 at 11:00 am.
ms660 is offline  
Old 22nd Jul 2018, 1:35 pm   #34
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

Andy, further to Post#33...

Don't forget that you can use the graphical method for finding ra without using a tangent or rough guessing, calculate the amplification factor (u) and the mutual conductance (gm) then ra = u/gm, as an example using the 6Q7 curves I used earlier, using the same operating point (250 Va & -3 Vg) as a reference:

For u, a change of grid voltage of 0.5 volts either side of 3 volts is a change of 1 volt, that plots along the constant anode current line (Ia = 1mA) as a change in anode voltage of 70 volts (approx.) so u=70/1 eg: u=70.

For gm a change of grid voltage of 0.5 volts either side of 3 volts is a change of 1 volt, that plots a change in anode current along the constant anode voltage line (Va = 250v) of 1.2mA (approx.) so gm=0.0012/1 eg: gm=1.2mA/V

ra=u/gm = 70/0.0012 = 58.33k which is pretty much spot on compared to the published figure of 58k in the valve data.

If you want to measure with a meter etc keep Vg constant, change Va, note the change in Ia....then ohms law....change in Va/change in Ia = ra.

Lawrence.
ms660 is offline  
Old 22nd Jul 2018, 1:43 pm   #35
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Cathode resistors and bypass caps.

Thanks Lawrence. I re-calculated parameters this morning and got for circuit - ra 42k, gain is (taking into account unbypassed Rk) = 34, gm = 1.1A/V. Actual gain for 0.5v input before clipping = 58? (OP of 29v RMS) . So my sums are off.

It's more the frequency and distortion I'm interested in at present but I probably need more practice and will look at yours results later Lawrence. ra is crucial for the accuracy or most other calculations as its used in them all.

Anyhoo, tested 1) un-bypassed (Un), 2) bypassed (By) by a 22u. The frequency response was better for the By = 3dB down at 12khz instead of 10khz for Un, but distortion was significantly worse = see pic 1 - Un is the blue trace, By is the orange. THD By = 4%, Un = 0.6%

I'm now going to test circuit a few config's which is 3) using a diode/s and 4) partial bypassed. 5) will be with a 0.1u cap and I might try some local FB and bootstrapping.

BTW what did Leak and Quad manage frequency response wise for their amps? Pic's 2 & 3 are my test setup.

Andy.
Attached Thumbnails
Click image for larger version

Name:	DSC00816.jpg
Views:	89
Size:	91.8 KB
ID:	166388   Click image for larger version

Name:	DSC00814.jpg
Views:	59
Size:	107.1 KB
ID:	166389   Click image for larger version

Name:	DSC00815.jpg
Views:	72
Size:	125.7 KB
ID:	166390  
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 22nd Jul 2018, 1:56 pm   #36
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by Diabolical Artificer View Post
... BTW what did Leak and Quad manage frequency response wise for their amps? ...
The frequency spec for the Quad II is 'within 0.2dB from 20Hz to 20kHz' but they don't say at what power (wide response is easier to achieve at lower powers not least because of the output transformer nonlinearities). Note Keith Snook's comments here http://www.keith-snook.info/quad-ii-...amplifier.html.

The Leak TL/12 Point One - Leak's first 'hi-fi' amp if you like, launched in the late 1940s - also claimed 20Hz-20kHz at +/- 0.1dB. I measured a Stereo 20 in tip-top condition which, at 1W output, was 0.2dB down at 20Hz and 0.5dB down at 20kHz, both with respect to 1kHz.

You'll struggle to find a speaker which is anywhere near that flat, of course.

Cheers,

GJ
__________________
http://www.ampregen.com

Last edited by GrimJosef; 22nd Jul 2018 at 2:07 pm.
GrimJosef is offline  
Old 23rd Jul 2018, 10:30 am   #37
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Cathode resistors and bypass caps.

That's some spec to aspire to G, albeit if the test is a tad fixed.

Tried the LED instead of resistor just now, f response slightly better - 3dB down ay 15khz but OP is way down 16v RMS and I had to reduce the IP V, I think there is blocking going on, see pic. Top trace is OP bottom trace is grid. THD is worse too @ 5.6%.

Can't quite get my head round this at present, need to read up, but I wonder if there is enough current through the LED, with 1.5v bias that's about 1mA. I tried a smaller mini LED, same result.

Andy.
Attached Thumbnails
Click image for larger version

Name:	DSC00819.jpg
Views:	68
Size:	87.7 KB
ID:	166424  
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 23rd Jul 2018, 10:43 am   #38
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Cathode resistors and bypass caps.

Looks like too much input for the bias.

Lawrence.

Last edited by ms660; 23rd Jul 2018 at 10:48 am. Reason: correction
ms660 is offline  
Old 23rd Jul 2018, 10:54 am   #39
GrimJosef
Dekatron
 
Join Date: Sep 2007
Location: Oxfordshire, UK.
Posts: 4,310
Default Re: Cathode resistors and bypass caps.

Quote:
Originally Posted by Diabolical Artificer View Post
That's some spec to aspire to G, albeit if the test is a tad fixed ...
Well, as I said I'm afraid, they paid qualified engineers to sweat over these problems for a lot longer than you or I can spare. Perhaps we shouldn't be surprised if, starting from scratch, we can't do as well as they did.

I've seen quite a few modern commercial valve amps and there are cases where the designer has lifted pretty much the entire small-signal circuitry from one of the classic amps. The alternative is a bit like designing your own carburettor (remember those ?). It's not impossible, but you'll be lucky to come up with one that works as well as Zenith's.

Cheers,

GJ
__________________
http://www.ampregen.com
GrimJosef is offline  
Old 23rd Jul 2018, 12:19 pm   #40
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,636
Default Re: Cathode resistors and bypass caps.

Scrub that last result, couldn't work out why the grid input Z had altered so much so I got out a 10k R to use as a grid blocker, then noticed I'd left my other low THD sig source plugged in to the grid. Took it out and bobs your uncle everything better.

Results for no Rk, replaced by three 1N4148's bypassed by a 100n cap are very good. F response a lot better, it's flat to 18hz, 1/2 dB down at 10hz. HF response better too, 3dB down at 25khz, however THD is 3.5%.

I'll have a look on a spectrum analyser to see how much is 2nd, 3rd harmonic etc. I might be able to get the THD down by using a bigger Ra and more HT and it may be worth putting up with a tad higher distortion on the front end for a very much improved f response.

I take your point G about the time etc and about lifting a proven design however I tried that approach on my first amp and got nowhere, also it's less fun building an existing design.

I'm away to read up on partial bypassing, bootstrapping and other methods.

Thanks for the continuing interest, Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 5:04 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.