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 > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 7th Jun 2020, 8:39 pm   #1
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Parasitic stopper: grid or anode?

I am commissioning my valve receiver project. I'm struck by the large number of tuned circuits you need with valves as compared to a solid state design. It's just a consequence of the higher impedances. I've attached a photo of the underside showing the myriad of trimmers. And yes, it's not beautiful but it *is* a development project. At least that's my excuse.

Needless to say, the receiver doesn't perform quite as expected.

I am using an E88CC double triode product detector. The 455KHz IF is fed to the grids in push pull and the audio is taken from the anodes also in push pull. A 3V p-p BFO signal is fed single ended to the commoned cathodes. If I look at the BFO input with a 150MHz scope I see "fuzz" on the rising edge of the 456KHz sine wave. A circuit simulation suggests stray inductance and capacitance in grid and anode circuits is causing VHF instability. If I add, say, 10 ohms resistance in series with both anodes or in series with both grids the parasitic goes away in the simulation.

Which is considered best practice: added resistance in the grid or in the anode circuit? Or both? Or doesn't it matter much?

73, Alan G3XAQ
Attached Thumbnails
Click image for larger version

Name:	IMG_2539.jpg
Views:	166
Size:	130.9 KB
ID:	207968  
Alan_G3XAQ is offline  
Old 7th Jun 2020, 9:02 pm   #2
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
Default Re: Parasitic stopper: grid or anode?

Decouple the anodes to ground, with something like 500pF.

Pass the audio through another filter: a 'pi' section with the outer elements being 500pF to ground and the linking element being 10KOhm.
G6Tanuki is offline  
Old 7th Jun 2020, 9:19 pm   #3
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

small signal stage in terms of power so stoppers could go anywhere. Large signal in terms of noise floor, so stoppers could go anywhere.

In general terms, with a three terminal device, you can have one terminal as a low impedance at RF (possibly wire length stray inductance is the problem) but not two or three.

In a differential pair product detector, the cathode-to cathode connection counts as low Z because the opposite valve cathode follows whatever drives its grid. So ideally you want a lossy impedance broadband on the grids and the anodes. 1/gm though says the cathode connection isn't too bad

Anode stopping is less damaging on low noise stages, grid stopping is less damagiing on efficiency of power stages, though they can be devils, You'll often see resistors inside small coils in anode connections of HF PAs.... designed to spoil impedances at VHF and up.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 7th Jun 2020, 9:24 pm   #4
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Parasitic stopper: grid or anode?

I already have 10n MLCC capacitors at each anode. I don't understand how a filter on the audio output will affect parasitics. Can you give some more detail? I've attached a circuit diagram to aid discussion.

73, Alan
Attached Thumbnails
Click image for larger version

Name:	pd.jpg
Views:	142
Size:	102.5 KB
ID:	207977  
Alan_G3XAQ is offline  
Old 7th Jun 2020, 11:23 pm   #5
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

10n to deck looks like a nice lowpass filter considering the Zout of the anode and the load resistor which effectively shunts it (and the zin of the next stage).

But at higher RF it looks like a dead short... only that connection has path length, and that makes nanohenries.

Make a low Z (at RF ) connection elsewhere, with a bit of series L and thet can get gyrated into a negative resistance... capable of cancelling the losses in some resonsnce or another - and then some.

Forget audio. You have to make sure stages are stable at all frequencies. Fail to do this and it's going to take advantage of the oversight. It may be oscillating at a silly frequency but it will exhibit side effects disturbng the stuff you want to handle.

This is a common oversight. Someone does a design, maybe produces Smith charts and s-parameter sets and Linvill stability criterion numbers, but this is usually seen through only the intended frequency range of operation. If you want a good circuit, there is no such limitation on the need for stability, you want that DC to daylight.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 8th Jun 2020, 6:33 am   #6
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
Default Re: Parasitic stopper: grid or anode?

No expert on RF but I've found grid stoppers more effective than anode stoppers. I've also found an RC attenautor from HT to anode better at stopping parasitic oscillation, determine osc freq, use 1/ ( 6.28 x R x C ) to find R/C values.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is offline  
Old 8th Jun 2020, 11:09 am   #7
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Parasitic stopper: grid or anode?

In post #3 David suggests grid or anode "stopper" resistors are both acceptable. In post #5, having seen the circuit, you seem to be hinting that an anode stopper might be the better choice but that a fully paid up RF designer would have recognised the intrinsic instability of the circuit and taken it into account, so avoiding the problem in the first place.

I can see that the anode stopper is added to swamp the reactance at VHF of the inevitably long leads between the valve pins and the audio transformer although I don't see how the value of resistance could be chosen any way other than "select on test". But I and probably other readers are itching to know if there is more to designing this product detector for VHF stability than just adding stopper resistors as bandaids during commissioning.

Oh, and are lossy ferrite anti-parasitic beads (3B1 ferrite) likely to be as effective as resistors? Maybe they just add more unknowns but they would be more convenient to use given the somewhat cramped space I am working in.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 8th Jun 2020, 12:48 pm   #8
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

Quote:
Originally Posted by Alan_G3XAQ View Post
I don't see how the value of resistance could be chosen any way other than "select on test".
That's pretty much how it's been done since time immemorial.

With modern simulation software you can analyse it.... well you would be able to analyse it if there were accurate models of the devices into RF frequencies - which there aren't. So stage 1 of the analysis process would be to gather a statistical sample of devices, build test fixtures and get going with a network analyser, then try to concoct equivalent circuits to model what you've just measured.

So most people simply have some favourite values they pulled out of thin air/copied from someone else and have proven themselves over time.

Beads do seem to work quite well - provided the oscillation frequency is high enough that the losses of that ferrite material have come to the fore. Fair-Rite inc. got a head start in this business by speccing their stuff not only for inductor/transformer use, but also as damper/absorber devices. So you can see in their data where beads start looking essentially resistive, and how many ohms.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 8th Jun 2020, 2:14 pm   #9
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Parasitic stopper: grid or anode?

OK, I've added 5x4x2mm ferrite beads at the anode pins on the product detector and the VHF "fuzz" I was seeing on the scope has gone. I've done the same on the front end mixer, which uses the same circuit topology, just to be sure.

The receiver sounds OK now but is a bit deaf. Some of the IF tuned circuits need more adjustment range than their trimmers provide. So onward and upward.

Thanks for the help.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 8th Jun 2020, 2:31 pm   #10
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,061
Default Re: Parasitic stopper: grid or anode?

Coming late to this party, but yes, exactly as described. Inspired trial is the way.

To be thorough, you'd try a shorter bead, of the same material, just to confirm the squirkles are strill suppressed, and then stay with what you've got. If not, you'd go up to a larger bead, so you know there's some margin in-hand.

Personally, I'd have started with grid stopper resistors. But, you have a solution, so full marks!
kalee20 is offline  
Old 8th Jun 2020, 2:35 pm   #11
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

It's quite possible that a lot of people building things don't have stoppers in place and get bursts of parasitics that they are completely unaware of. If they'd looked with a scope and had the right suspicious mindset, they'd have found them.

Put down careful stopping as good housekeeping and hygiene.

At the interview for my current job, we were all sat around a round table and I was passed a board, the transmitter for a new transponder. I asked a few quastions about what it was, power, frequency etc. And "Where are the stoppers?" the reply was "The What?"

"They're normally needed to prevent spurious oscillation. How've you managed to do without them?"

Guess what the problem was they'd been having with the board? Guess who'd just got a job.

It turned out the gain per transistor was miserable. 6.2dB in the PA on the transistor data sheet. The drivers were hardly any better. Efficiency about 30% but as tough as old nails. Turned out there was no margin in the needed gain for proper stoppering, so getting it sorted was a bit of a knife-edge job. Serious wile and the occasional tweak in production was needed. Since then I've done a few varieties of new designs for different sets. All have plenty of gain, good stoppering and easy margins in production. Much to production's delight.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 8th Jun 2020, 4:05 pm   #12
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Parasitic stopper: grid or anode?

I only have (extremely) limited experience of valves/tubes but when I played with a 12AU7 a while back I was able to model the small signal behaviour quite well up into VHF. This took maybe an hour of effort to derive a VCCS based model of the valve in its base. I had to include the effect of the dielectric of the base material as this causes extra capacitance between the pins. I also included a guess at the inductance of each electrode within the valve itself.

I used the datasheet capacitance between electrodes and chose the Gm based on the datasheet. See post #71 in this thread.

https://www.vintage-radio.net/forum/...=161894&page=4

I didn't try very hard and I suspect that it's possible to model these valves quite well up to about 300MHz. Maybe the model will be a bit flawed but I've always maintained that if you can't/won't make a basic model of a system then you can't claim to understand how the system works or how/where it may go unstable. All you can do is prod it and see what happens when you make changes to the circuit. When it eventually stops going unstable then I guess you can call it a successful outcome.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 8th Jun 2020, 5:14 pm   #13
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

One advantage to valves is that there are relatively small spreads in parameters if you've been used to semiconductors, and there aren't opportunities for future production to miraculously become ten or more times faster.

So once you've found what works as a stopper and gone just a bit further, it will probably be fine in production and for the future.

People making the transition from valves to semiconductors had it much harder.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 8th Jun 2020, 8:13 pm   #14
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Parasitic stopper: grid or anode?

So far I've not had a problem when trying to create small signal models based on the 12AU7 datasheet. I only have a 12AU7 valve here to play with and I dug it out this evening and put a 47R metal film resistor in the cathode and 6800R at the grid and 3900R at the anode. I also added these components to the basic model based on the datasheet and added a suitable amount of lead inductance for each connection. The only 'clever' or unique thing about the model (that isn't on the datasheet) is the extra pin to pin capacitance between a-g and g-k that is caused by the valve base.

See below for a measurement in the input impedance at the grid when measured on a VNA and also the input impedance predicted by the model. They seem to agree very closely.

In both plots I have cheekily added about 10nH extra inductance at the input and this shifts the resonance to a point just over 250MHz as this is where the resistance dips negative. So this creates the conditions for oscillation.

I then added this shunt inductance to the input of the real 12AU7 circuit and sure enough, it oscillated at 260MHz as in the spectrum analyser plot below. It was spot on with the model!

The product detector circuit will be a bit different because the valves will be driven such that Gm varies a lot across the LO RF cycle. So the valve will presumably be more prone to brief bursts of instability over part of this cycle if there is little or no margin against instability.

Anyone can create the model I used here in just a few minutes if they read the datasheet for the valve.
Attached Thumbnails
Click image for larger version

Name:	12AU7_model_VNA.jpg
Views:	56
Size:	73.0 KB
ID:	208052   Click image for larger version

Name:	12AU7_260MHz.gif
Views:	51
Size:	15.2 KB
ID:	208053  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 8th Jun 2020, 9:00 pm   #15
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Parasitic stopper: grid or anode?

Yes, Jeremy, the product detector burst into VHF oscillation only during about half the rising edge of the cathode drive, so as the triodes were turning off. I'm using an E88CC double triode so the gm is somewhat higher than your 12AU7 experiments. I never measured the oscillation frequency and I've "cured" the instability with the ferrite beads I mentioned but I'd still be interested in knowing what software you used for your modelling. Your thread reference to a post #71 doesn't tell the inexperienced reader how you produced the plots.

73, Alan G3XAQ
Alan_G3XAQ is offline  
Old 8th Jun 2020, 10:15 pm   #16
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Parasitic stopper: grid or anode?

I used an old works copy of Eagleware Genesys. This is really old software now as it is V2004 and was the last version before it was taken over by Agilent.

Out of the box the Genesys software can't produce these plots very easily, I have to use the Genesys equations/maths engine to post process the reflection coefficient data and this lets me plot it as Rs+Cs/Ls or Rp+Cp/Lp etc.

It can also post process the raw VNA data and plot these directly alongside the above. This makes it easy to make comparisons.

I think LTSpice can do things like this but it may take a fair bit of keyboard bashing to configure it all.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 8th Jun 2020, 10:53 pm   #17
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Parasitic stopper: grid or anode?

Quote:
I'm using an E88CC double triode so the gm is somewhat higher than your 12AU7 experiments.
I created a crude model of your valve and it looks like this valve can generate a lot of negative resistance over a wide bandwidth. I think this initial 'page 1 of the logbook' model would be reasonable up to maybe 300MHz but it would be wise to compare it against a few VNA measurements and give it a few tweaks. It didn't take much resistance in the anode to get K to go over 1 but a lot depends on the layout and the lead inductance in the 10nF cap at the anode.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 8th Jun 2020, 11:15 pm   #18
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,801
Default Re: Parasitic stopper: grid or anode?

Quote:
Originally Posted by G0HZU_JMR View Post
and it looks like this valve can generate a lot of negative resistance over a wide bandwidth.
Words to send me running away, quickly.

It's got the internal inductances of its lesser cousin plus more gain. So it's not unexpected.

Time-of-flight effects are difficult to model. Electron velocities are low and distances are mundane, but capacitance coupling is at close to 100% of C. This also makes life difficult. Velocity isn't constant s=t*(v+u)/2 and the mean value of u is zero. v is from anode volts *e = (1/2)*m*(v squared)

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is offline  
Old 8th Jun 2020, 11:38 pm   #19
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Parasitic stopper: grid or anode?

Having spent a few minutes playing with the model of the E88CC (in a modelled valve base) it reminds me of the J310 JFET in terms of the negative resistance and capacitance it produces vs frequency. The capacitances and the Gm are similar too. If it can oscillate up around 400MHz then it would be harder to model this because of the effects of the valve base and other effects as you describe. I found that my previous valve models started to misbehave badly above about 330MHz.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 9th Jun 2020, 10:30 am   #20
Alan_G3XAQ
Pentode
 
Join Date: Apr 2018
Location: Canterbury, Kent, UK.
Posts: 189
Default Re: Parasitic stopper: grid or anode?

Hmm. I chose the E88CC partly because I had a pair in the junk box and partly because the high gm of the frame grid construction seduced me into thinking it would make a good switching type mixer when driven with two or three volts p-p of oscillator. Changing to a 12AT7 would need surgery to the heater wiring so I'm not enthusiastic for change.

On the other hand, a ferrite bead at each anode appears to have tamed it, at least as high as I can measure with my limited test gear, up to maybe 200MHz. The old beads in the junk box are almost certainly Philips 3B1 ferrite, which is similar to Fair Rite 43. A 2661000301 bead has over 50 ohms loss resistance above 50MHz.

Perhaps in a future life I will steer away from "sexy" valves and choose the more mundane workhorses.

73, Alan
Alan_G3XAQ is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 6:38 am.


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.