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 > Specific Vintage Equipment > Vintage Tape (Audio), Cassette, Wire and Magnetic Disc Recorders and Players

Notices

Vintage Tape (Audio), Cassette, Wire and Magnetic Disc Recorders and Players Open-reel tape recorders, cassette recorders, 8-track players etc.

Closed Thread
 
Thread Tools
Old 13th Nov 2020, 4:53 am   #1
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Ferrograph Series 4 decoupling capacitor values

I've got motorboating when I increase the volume on the Ferrograph, and when it occurs the motor solenoid releases. I put this down to an extra current draw causing a B+ drop which de-energises the solenoid.

The anode voltages are high on the amplifier stage. I've removed the two dual-can 8µF+16µF capacitors on the amplifier chassis that decouple V1 and V2 (C9-C11 & C12-C13). Shorting the second input jack between V2 and V3 stops the motorboating so it must be around V1 and V2. I homed in on the dual cans because I haven't replaced them, and in my logic leakage there could increase the B+ current as well as causing the motorboating.

Perhaps I'm completely wrong, but if not I'd like to congratulate myself on my rudimentary troubleshooting.

I'm getting 6mA leakage before reforming, so I might be in the market for some new capacitors. I'm confused because I was checking that a new dual 16µF can would be suitable, and see that C9 (8µF) is the screen bypass capacitor for V1 (EF80). Performing the calculation to make this a high-pass filter for audio frequencies [Cg2 = 1 / (2 pi f Rg2)], with a screen grid resistor of 270k ohms it seems 47nF would be more suitable!

What's going on here? Could I put two new dual 16µF cans in place instead without messing up any filters?
Attached Thumbnails
Click image for larger version

Name:	Ferrograph 4A Amplifier Stage 1.PNG
Views:	72
Size:	196.9 KB
ID:	220247  
Uncle Bulgaria is offline  
Old 13th Nov 2020, 8:46 am   #2
Richardgr
Heptode
 
Join Date: Oct 2019
Location: Täby, Sweden
Posts: 692
Default Re: Ferrograph Series 4 decoupling capacitor values

I think you have a moderately regulated screen supply, and there is no signal in that path. The signal goes to the grid via transformer coupling.

So it is just a decoupling capacitor.
Richardgr is offline  
Old 13th Nov 2020, 11:09 am   #3
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

As said, it's the de-coupling capacitor for g2 of V1, its reactance at 100Hz is approx. 199 Ohms, so no problem.

Lawrence.
ms660 is offline  
Old 13th Nov 2020, 2:14 pm   #4
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

Won't there be signal by the nature of the pentode? As this grid is the screen grid, I understand its effect is to minimise the Miller capacitance by maintaining a steady charge. Surely electrons are still hitting the screen grid no matter whether the valve is coupled capacitively or inductively.

That makes sense to me, Lawrence - so why do the valve design sites I've been consulting (Lenard Audio, Valve Wizard, R-Type etc.) use that formula which for 270k ohms results in very small capacitances?

Doubling it here to 16µF, as a widely available new dual capacitor, would reduce the reactance so would be an improvement, I suppose.
Uncle Bulgaria is offline  
Old 13th Nov 2020, 2:33 pm   #5
wd40addict
Octode
 
wd40addict's Avatar
 
Join Date: May 2009
Location: Stevenage, Herts. UK.
Posts: 1,515
Default Re: Ferrograph Series 4 decoupling capacitor values

Dodgy decouplers leading to motorboating is a stock fault on Ferrographs.
wd40addict is offline  
Old 13th Nov 2020, 2:39 pm   #6
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Re: Post#4: The capacitor is there to prevent/minimize any signal from whatever source appearing on g2, a signal on g2 would modulate the cathode to anode electron stream.

If you do the potential divider maths for R1 = 270k and R2 = 199 Ohms (reactance of the capacitor at 100Hz) you can see the magnitude relationship between a signal on g2 with respect to the signal as presented across R1 and R2.....Eg: If there was say an AC voltage of 1 Volt at 100Hz across the that potential divider (1 Volt AC on the HT rail) then the magnitude of that AC voltage as presented to g2 would be approx. 736uV....not a lot considering that g2 is much further away from the cathode than g1 is and therefore any modulation effect on the electron stream is much less.

Remember that all circuits are potential dividers.

Lawrence.

Last edited by ms660; 13th Nov 2020 at 2:48 pm. Reason: addition
ms660 is offline  
Old 13th Nov 2020, 8:19 pm   #7
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

Apropos of your voltage divider comment, Lawrence: the R-type 'rule of thumb', quoted in their Graham Dixey articles, is for screen grid capacitors to be 1/10 or less of the screen grid resistor, so it can get very big before approaching 1/10 of 270k! The equation suggestions fractions of a microfarad would be ideal, but even 16µF is only 497 ohms at 20Hz.

Judging from the calculation, increasing the capacitance in the other decouplers can only benefit the stability of the B+. What is the disadvantage of this approach?

One of the dual cans has dropped to µA leakage after reforming each side, so they may be salvageable.
Uncle Bulgaria is offline  
Old 13th Nov 2020, 9:45 pm   #8
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Quote:
Originally Posted by Uncle Bulgaria View Post
Apropos of your voltage divider comment, Lawrence: the R-type 'rule of thumb', quoted in their Graham Dixey articles, is for screen grid capacitors to be 1/10 or less of the screen grid resistor, so it can get very big before approaching 1/10 of 270k! The equation suggestions fractions of a microfarad would be ideal, but even 16µF is only 497 ohms at 20Hz.
So according to your quote from the article you mentioned, the reactance of the screen grid de-coupler at a given frequency could be less than 1/10th of the resistance of the screen feed resistor.....how's about 1pF then at 100Hz?

If replying to me in the future please bear in mind that I left skool at 15 years of age with no certificates/qualifications whatsoever so fancy words go right over my head to the point that I can't be bothered to look up what they mean....!

Lawrence.
ms660 is offline  
Old 13th Nov 2020, 9:57 pm   #9
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

Apropos of my quote? The best thing about words is the mouthfeel. Some feel good when they come out, some don't! I like that one.

1pF? I have absolutely no idea! I'm trying to understand the parameters here. Skool or no, I have zero qualifications in circuitry and every problem and investigation throws up numerous questions. Presumably they chose two 8+16µF capacitors for a reason; perhaps cost and reducing the number of different parts.

I'm confused because that article, and others, suggest a calculation for that capacitor which results in values magnitudes lower than that included by the circuit. I realise that the higher values are still low reactances at 100Hz (or the 20Hz of the article) but it's interesting they used one so far away from ideal.

I've ordered a couple of dual 16µF capacitors so I'll see how they go. One of the cans has a half that, despite low leakage, had a capacitance measured in pF...on second thoughts, perhaps I'll make it the screen capacitor!
Uncle Bulgaria is offline  
Old 13th Nov 2020, 11:44 pm   #10
DMcMahon
Dekatron
 
DMcMahon's Avatar
 
Join Date: Apr 2019
Location: Worthing, West Sussex, UK.
Posts: 6,587
Default Re: Ferrograph Series 4 decoupling capacitor values

Out of interest Uncle Bulgaria, how are you measuring/determining your capacitor leakage current values. Is this something you can do with your EPE tester ?

David
DMcMahon is offline  
Old 14th Nov 2020, 12:28 am   #11
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

I'd imagine it's from nothing else, David! Yes, the EPE tester does that. One sets the voltage and application time, then as it runs it reports applied voltage and leakage current, updated a few times a second. Who knows how accurate it is? But it's something!

I've got the capacitance values from one of those 'component tester' units which are built from that German chap's design which has a large following over at the EEVblog. It reports capacitance and ESR. Again, I don't know its accuracy, but between them the two seem to flag the duff ones.
Uncle Bulgaria is offline  
Old 14th Nov 2020, 11:00 am   #12
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Ignore the 1st paragraph in Post#8, I was getting the values mixed up The edit time had run out and it was approaching sleep time for me, hence no notice for that error yesterday.

Lawrence.
ms660 is offline  
Old 14th Nov 2020, 11:53 am   #13
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Uncle B, if you haven't got the detail but want the detail further to what's quoted in the rtype org article then scroll down to Section 12.3 (iii) in the link below (book page 496 of the Radiotron Designers Handbook):

http://www.tubebooks.org/books/rdh4.pdf

Lawrence.
ms660 is offline  
Old 14th Nov 2020, 3:11 pm   #14
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

I see I've erred too in saying V1 was an EF80 when it's an EF86 way back in post #1.

I've been puzzling over that section you've kindly posted, in relation to the EF86 datasheet as a way of generating the screen capacitor value, but have got stuck on 'm', or Ib/Ic2 which is described only as 'assumed constant'. I can't find a reference to c2, though I assume Ib is the bias current. It's a great book, but rather too advanced for me. Thus demonstrating the usefulness of a maths A-level. I couldn't derive m from a previous section as it doesn't appear to be present.

I'm just finding it hard to understand why this capacitor value is nominally calculated at fractions of a µF, but it appears there'd be negligible difference at audio frequencies if it were replaced with a capacitor larger by a couple of orders of magnitude.
Uncle Bulgaria is offline  
Old 14th Nov 2020, 3:41 pm   #15
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Quote:
Originally Posted by Uncle Bulgaria View Post
I've been puzzling over that section you've kindly posted, in relation to the EF86 datasheet as a way of generating the screen capacitor value, but have got stuck on 'm', or Ib/Ic2 which is described only as 'assumed constant'. I can't find a reference to c2, though I assume Ib is the bias current. It's a great book, but rather too advanced for me. Thus demonstrating the usefulness of a maths A-level. I couldn't derive m from a previous section as it doesn't appear to be present.
m = Ib/Ic2.....where:

Ib is the anode (aka the plate) current.

Ic2 is the screen grid current.

"assumed constant" means that the ratio obtained from the above calculation is assumed to be constant.

If you look at the valve data (6J7) for the example given (in the book) in the link below and scroll down to Typical Operating Conditions and Characteristics you will see the typical characteristics for that valve configured for pentode operation, there you can see that the anode current is given as 2mA and that the screen grid current is given as 0.5mA, dividing the former by the latter gives a figure of 4 which is the same figure (m = 4) that's given on page 497 in the book:

http://www.r-type.org/pdfs/6j7-1.pdf

Lawrence.
ms660 is offline  
Old 14th Nov 2020, 3:56 pm   #16
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

The high value screen grid de-coupler for V1 might be to reduce the level of any hum in the screen grid circuit, don't forget that V1 is the first stage in a triple stage voltage amplification chain when it's used for the input, that's all I can think of at the moment.

Lawrence.
ms660 is offline  
Old 15th Nov 2020, 4:55 am   #17
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

That's cleared the fog! Thank you. I had got the idea it was a universal constant like the speed of light 'c'.

The EF86 datasheet had the more intelligible (to me, at least) Ia for anode current, and Ig2 for the screen grid.

After muddling my units (the EF86 gm was given in mA/V so my nominator was far too large) I ended up with an m of 5, so the B = 1+Rs.gm/m.µt = 3.56.

Thus Cs = B/π.f.Rs = 0.2µF for 20Hz, or 0.04µF for 100Hz. Bigger is better as far as approaching DC is concerned? Rearranging for f gives 0.5Hz for 8µF. Wouldn't it be best to have a higher cutoff if hum were the concern? For argument's sake you'd lose a lot of audio, but a smaller value would reduce gain at 50Hz, if I've interpreted the graph on p497 correctly.

I don't understand the 'limiting loss of voltage gain' being 1/B = 0.06 = 24.3dB, as I don't understand decibels, but I guess that it means there's a lower limit derived from the valve's working parameters beyond which voltage gain is impossible.
Uncle Bulgaria is offline  
Old 15th Nov 2020, 4:05 pm   #18
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

Quote:
Originally Posted by Uncle Bulgaria View Post
I don't understand the 'limiting loss of voltage gain' being 1/B = 0.06 = 24.3dB, as I don't understand decibels, but I guess that it means there's a lower limit derived from the valve's working parameters beyond which voltage gain is impossible.
Here's a site worth saving for dB calculations:

http://www.sengpielaudio.com/calculator-gainloss.htm

For voltage gain or loss to dB's use the second calculator on that page.

With regards to the graph on page 497 and B = 16.5 more sense of the quoted 24.3dB limiting loss figure can be seen if you extend the logarithmic frequency scale another decimal place shift to the left down to 0.5 Hz then align the frequency that's derived from Rs/Xs = 1 (approx. 1.5Hz) with 1 on the Rs/Xs scale, the limiting loss (use the 0.5 Hz mark) as quoted will be in near enough in agreement with the graph as should the quoted 1dB loss at 50 Hz as read off from the computed value of Rs/Xs at that frequency on the Rs/Xs scale (computed value approx. 33)

The losses are due to the increasing de-generation which results with an increasing reactance of the de-coupling capacitor as the frequency is reduced.

Remember that the screen grid voltage has an influence on the anode current, if it dips then so does the anode current as does the voltage developed across the anode load.

Lawrence.

Last edited by ms660; 15th Nov 2020 at 4:13 pm. Reason: words change
ms660 is offline  
Old 19th Nov 2020, 12:56 am   #19
Uncle Bulgaria
Nonode
 
Uncle Bulgaria's Avatar
 
Join Date: May 2016
Location: Cornwall, UK.
Posts: 2,315
Default Re: Ferrograph Series 4 decoupling capacitor values

Once again you come up trumps, Lawrence! That's going to be very good for getting my brain around the concepts when it comes to gain and loss comparisons. I've traced off the scale and using it in conjunction with the graph makes some of the mists dissipate.

I'm going to reference this when doing the CR 240 - I haven't got my reels with me so I can't set up the bias and go through all the checks on the Ferrograph yet.

I've replaced the capacitors with the dual 16µF ones I'd ordered, and replaced a few more out of tolerance old resistors. A sine wave in at J1 is pristine at the loudspeaker output. I'm impressed. It's playing some music through J2 now through the internal 'speaker and sounds fantastic.

The valve voltages were all about right. Cathode voltages were very close, but anode and screens were high by about 20V, probably because the HT is higher than nominal.
Uncle Bulgaria is offline  
Old 19th Nov 2020, 11:21 am   #20
ms660
Dekatron
 
ms660's Avatar
 
Join Date: Apr 2011
Location: Cornwall, UK.
Posts: 13,454
Default Re: Ferrograph Series 4 decoupling capacitor values

No problem, once you get a basic understanding of valves things get easier to understand, getting to grips with the various symbols and definitions when reading different books and and different valve data sheets can sometimes be confusing at first but those differences usually refer to the same thing.

Lawrence.
ms660 is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 9:44 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.