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 2019, 5:26 pm   #1
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
Default Valve preamp.

After rootling out some bits I decided to have a go at knocking up a preamp to go with the big amps I made. They need 1.5v RMS to get the full 100w OP, my PC and other music sources do a bit less than this, so the preamp will compliment them.

At first I was thinking of designing the circuit myself, maybe using EF80's triode strapped for a line level amp, but after some searching online I decided to use Patrick Turners design's for the phone and line level amp, the fettling will take enough time as it is, so decided to use an existing design. I have a lot of respect for his work, he put in a lot of research and testing and to be honest a phono amp is a complex beasty, anything I designed wouldn't be half as good as his.

I've had this big tfmr and choke knocking about for ages, see pics. The tfmr is huge in comparison to it's power output, only 80mA, so perfect for a preamp. I've connected it with an LC filter rather than CLC, to reduce the HT, with two 820u 400v caps ripple is 20uV before regulation. I'll use Patricks PSU design as a guide and fit it what components I have.

Any ideas and constructive criticism welcome, Andy.
Attached Thumbnails
Click image for larger version

Name:	schem-linestage-2005-edited2011.jpg
Views:	485
Size:	87.6 KB
ID:	184612   Click image for larger version

Name:	schem-rocket-ph-amp-sheet-A-2011.jpg
Views:	395
Size:	107.8 KB
ID:	184613   Click image for larger version

Name:	schem-preamp-psu-2005-edited2011.jpg
Views:	340
Size:	114.2 KB
ID:	184614   Click image for larger version

Name:	schem-rocket-ph-regulators-sheet-B-2011.jpg
Views:	337
Size:	100.1 KB
ID:	184615   Click image for larger version

Name:	DSC01390.JPG
Views:	336
Size:	160.7 KB
ID:	184616  

__________________
Curiosity hasn't killed this cat...so far.

Last edited by Diabolical Artificer; 7th Jun 2019 at 5:33 pm.
Diabolical Artificer is online now  
Old 7th Jun 2019, 7:36 pm   #2
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,288
Default Re: Valve preamp.

The use of a constant current Anode load for the line amp is interesting. I considered using such an arrangement for a preamp I was designing but when I modelled it in Ltspice it had far too high an output impedance, even the 100k pot may be too great a load.

Peter
Electronpusher0 is online now  
Old 7th Jun 2019, 11:35 pm   #3
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Valve preamp.

The constant current source in the anode is completely wasted. The datasheet shows the anode curve resistance at that bias voltage to be around 7k Ohms, and then the 100k pot is in parallel with that. The voltage drop can be done at the marked 3mA with a 56k resistor... this is dominated by the anode curve resistance, and the transistor will only make a finite load Z not that much bigger than 56k.

It's easy to say "Ooh look it uses a triode" and "A constant current source is used to maximise the gain" and these statements will be swallowed whole by anyone who doesn't check the arithmetic.

Moving on to the second diagram, the use of a common-source JFET for a moving coil input is most peculiar. JFETs only give their best noise performance with much higher source impedances than MC cartridges present. This isn't just the wrong device, it is the wrong type of device. Maybe something could be done with an array of them in grounded gate mode, but this isn't that.

With this circuit someone can say "Oooh look, it doesn't use feedback!" and "It uses a passive RIAA network" these factors seem to give some people the warm and fuzzies, but just being able to point to these features on a schematic is no indication that the circuit is properly designed.

Sorry, Andy, but there are aspects of those circuits which tick trendiness boxes but leave the engineering boxes blank.

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 2019, 8:22 am   #4
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
Default Re: Valve preamp.

"The constant current source in the anode is completely wasted....." Interesting, I knocked this circuit up and briefly tested it and thought it ok, it was flat from 10hz to 20khz. I did the CCS differently though using a zener to fix BE. Pat Turner just uses a few resistors to set base and CC, I've found that this doesn't perform well and prefer a preset as part of Re so I can adjust the CCS.

" the transistor will only make a finite load Z not that much bigger than 56k." If I use a two tranny CCS this should improve things yes? Higher Z.

Re using a FET, the whole article is here - http://www.turneraudio.com.au/preamp...hono-2005.html I'll admit to to not being able to follow some of PT's writting and explanations and what I know about FET's and RIAA curve and phono stages would fit in the pocket of a micro pixie.

Moving on I also looked at this one - http://www.valvewizard.co.uk/phonopcb.html see below, but it's for MM only, not switchable between MC and MM. Also looked at this one - http://www.dogstar.dantimax.dk/tubestuf/sapover.htm and lastly had a peak at Morgan Jones design, see below.

Finding a decent phono stage has been a bit daunting, 1) There's a lot of BS and spiel on the subject, 2) There are reams of technical brain sludge to wade through 3) I'm trying to find a design that uses valves/bits I have 4) I really can't be arsed to spend ages building, testing a stage I'm not that bothered about as my record player isn't an all singing all dancing jobbie and I don't play records much anymore.
Therefore any guidance welcome, thanks for running your eye over this for me David.

Andy.
Attached Thumbnails
Click image for larger version

Name:	Valve Wizard phono stage.jpg
Views:	195
Size:	70.4 KB
ID:	184656   Click image for larger version

Name:	Fred Nachbauer's phono stage.jpg
Views:	203
Size:	99.0 KB
ID:	184657   Click image for larger version

Name:	Morgan Jones Phono Stage.jpg
Views:	172
Size:	55.0 KB
ID:	184658  
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is online now  
Old 8th Jun 2019, 8:55 am   #5
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,288
Default Re: Valve preamp.

I attach my design for an RIAA preamp.
I built this for my home set up which uses Moving Magnet so the design does not account for MC.
I used passive equalisation and no overall feedback. The output cathode follower is used because the RIAA stage is a stand alone unit attached to the main amp by cable.
I have not got any tone controls, no reason other than I have not round to it yet.
I have built it and used it for some time. It sounds good to my untutored ears.

Peter
Attached Files
File Type: pdf Visio-RIAA Preamp.pdf (15.8 KB, 192 views)
Electronpusher0 is online now  
Old 8th Jun 2019, 9:19 am   #6
Electronpusher0
Nonode
 
Join Date: Jun 2016
Location: Bognor Regis, West Sussex, UK.
Posts: 2,288
Default Re: Valve preamp.

I should add that I used an excel spreadsheet to calculate the equalisation components (attached) and have plotted the actual frequency response on the bench, it follows the RIAA curve within the limits of my equipment.

Peter
Attached Files
File Type: xlsx RIAA calcs.xlsx (33.7 KB, 90 views)
Electronpusher0 is online now  
Old 8th Jun 2019, 9:59 am   #7
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Valve preamp.

Quote:
Originally Posted by Radio Wrangler View Post
Moving on to the second diagram, the use of a common-source JFET for a moving coil input is most peculiar. JFETs only give their best noise performance with much higher source impedances than MC cartridges present. This isn't just the wrong device, it is the wrong type of device. Maybe something could be done with an array of them in grounded gate mode, but this isn't that.
Well no. The resistive contribution to noise is the channel resistance, and this is approximately 2/(3gm). For a 10mA/V jfet gm that is about 70 ohms. That is about right - the low noise p-channel LSK74 http://www.linearsystems.com/lsdata/...%2010%2006.pdf has 0.9nv/rootHz - a noise resistance of 50 ohms.

Now it depends on what sort of MC cartridge you are looking at. But the Denon DL103 has a 40 ohm coil resistance, and matches very well with a single LSK74.

Now you can argue that bipolar transistors can have a lower equivalent noise resistance. However, good luck finding any of those, which the semiconductor industry has been hell bent on obsoleting in recent years. If it doesn't have a huge market, and eg go in a mobile phone, a PC or a flat screen TV they lose interest fast.

The first to go about fifteen years ago was the superb Rohm 2SB737 with 0.2nV/rootHz at 10mA, and an rbb' of 2 ohms. This is a trend that has continued year on year since.

Craig
Craig Sawyers is offline  
Old 8th Jun 2019, 10:43 am   #8
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Valve preamp.

Quote:
Originally Posted by Diabolical Artificer View Post

" the transistor will only make a finite load Z not that much bigger than 56k." If I use a two tranny CCS this should improve things yes? Higher Z.
Andy.
One of the two transistor constant current sources can indeed be made to be a better constant current source than a single transistor one buy it won't make much difference in this circuit because the anode of the valve creates a far lower impedance of its own.... about 7k according to the valve data sheet. This is not a resistor that gets drawn on a schematic, so it's easy to not see that it is there.

If I measure a constant current source, I'll find that the current value changes a bit as I sweep the voltage. Plotted, I get a slope on the current/voltage characteristic. Thevenin and Norton models represent this as an ideal source, plus a resistance. Using the current source as an anode or collector bias source in an amplifier, this resistance will reduce the intended load resistance, and will reduce the gain.

If I measure the anode current of a valve as I sweep its anode voltage, I also find that the current changes a bit. This puts a slope on the Ia/Va plots, right across the active region and my anode is behaving like it had an internal loading resistor. This shunts the intended load resistance and reduces the gain below what you'd expect from the schematic.

This 'Anode impedance' combined with the valve's Gm sets a maximum gain for the device even if all other loads were infinite. Gm*Ra is quoted on the datasheet as mu and for this valve is about 20. Have a look at the mu values for ECC81, ECC82, ECC83 and you'll see that the ECC83 is special. It was designed for high mu (for a triode) and it delivers. But what was done to achieve this confines it to lower powers and lower frequencies than its stable mates. Look at its anode curves and the ECC83 ones are less sloped. This is just another view of the same thing.

What goes on in the valve itself? The electric field created by the anode voltage is not negligible near the grid, and gets to influance the grid's control og the electron flow from the cathode.

The tetrode's screen grid lives up to its name. It is an electrostatic screen whose job is to protect the grid 1 area from the anode field. The result is a set of flatter anode curves at the price of a dip (the dreaded kink) where the anode voltage drops below g2 and g2 therefore looks like a more interesting target. The cure for this is a suppressor grid that screens the g2 area from seeing the reduced anode field, suppressing any electrons that overshoot g2 and would have decelerated and gone back. It discourages secondary emission electrons kicked off of the anode as well. These added grids give a much flatter anode curve set, and mu is much increased. This is the whole raison d'etre of tetrodes and pentodes for non-RF uses. At RF, the screening of anode to g1 capacitance is anothe string to their bow.

Now have a look at the mu figure for the EF86.

OK, EF86s can be microphonic and can f=have some noise disadvantages, but they are the entry into a world of higher stage gains BUT only so long as the other stuff presents a high enough Z to the anode.

In the schematic posted, the triode itself is the dominant limitation. Chucking the current source away and fitting a 56k resistor would reduce the gain, but only by a little.

Seeing the constant current source there will make people think that stage must give oodles of loop gain. But once you do the sums it turns out to be a mirage.

Triodes, pentodes, bipolars, FETs are ALL imperfect. The art of electronics lies in exploiting their respective strengths while simultaneously covering for their respective weaknesses.

Not doing the arithmetic leads to oversimplified beliefs that triodes, through being simpler, are always better than pentodes and leads eventually to beliefs like rotating valve-holders so that the Mullard logo is aimed at Blackburn will improve the sound

Going into details about the noise parameters of JFETs versos the source impedance applied to them shows that JFETs in common-source circuits are only a good choice for quite large impedances. Moving coil cartridges are amongst the lowest impedance sources you're ever likely to have to handle... and also the one most critical on low noise design. The device is diametrically opposite what you need for the task. You might be able to design a circuit to make up for this by having an immense number of them operated in parallel, but other limitations come into play. Fad and fashion and USPs sell audio equipment. Arithmetic designs good equipment - if it's not overruled. This is an environment where the bipolar starts with a useful advantage.

David

Edit - I've just read what Craig posted while I was typing.... yes, the best bipolars have disappeared. Mixing desk microphone amp designers are desolate.

I was involved in the noise measurement business for work-work and that went into making measurements to plot contours of constant noisiness on Smith charts of presented source impedance. This is do-able down to lower RF frequencies, but gets cumbersome below that. Anyway, the trends can be seen and extrapolated downwards. Active devices have several noise-generating mechanisms and different ones achieve significance in extreme applications. The trend is for Zoptimum for noise to move closer to the real axis and to go higher.

I suppose we're stuck simulating those low base spreading resistance bipolars with ayyays of parallelled common or garden varieties.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

Last edited by Radio Wrangler; 8th Jun 2019 at 11:03 am.
Radio Wrangler is offline  
Old 8th Jun 2019, 12:20 pm   #9
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Valve preamp.

Looking at the circuit, my first thoughts were, "Why does the cathode bypass 470μF capacitor have 1μF across it?" And "Why is there a constant-current source as the triode load?"

Modern electrolytics perform pretty well to high frequencies - there is no benefit in trying to bypass ESR or ESL with a small non-polarised component. It may be necessary if the component layout is really bad - if, for instance, C2 is located 10 feet away and connected to the rest of the circuit by widely-spaced leads - then local bypassing would help. But unless there is an audiophool reason for doing that, nobody's going to!

The constant-current load is, as David and Craig point out, going to give negligible benefit because the stage gain is already dominated by the ra of the triode itself and with the following grid loading resistance. Using a CC load might result in a 1% increase in gain (I've not done the calculations) but that's it. It might have been more elegant to use two series-connected resistors as the anode load, and bootstrap the junction to the cathode-follower output so that with volume control at maximum the bootstrapping boosts the anode load resistance- but even then the effective load is limited by the following grid resistors and volume control.
kalee20 is offline  
Old 8th Jun 2019, 4:08 pm   #10
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
Default Re: Valve preamp.

Thanks, lots to think about and digest, I'll respond when I've had a thunk.

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is online now  
Old 8th Jun 2019, 5:15 pm   #11
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Valve preamp.

Quote:
Originally Posted by kalee20
Looking at the circuit, my first thoughts were, "Why does the cathode bypass 470μF capacitor have 1μF across it?" And "Why is there a constant-current source as the triode load?"
As you know, the correct answer to this question is 'because certain people expect to see these features and wrongly regard them as signs of good audio design'. Unfortunately, these audiophile fads could degrade performance. Bypassing a capacitor can raise RF impedance, while leaving audio impedance largely unaffected. A poor CCS could add distortion through its own impedance nonlinearity.
G8HQP Dave is offline  
Old 8th Jun 2019, 5:26 pm   #12
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Valve preamp.

By the time you get to RF frequencies, the wound foil construction of a large value electrolytic capacitor can be looking somewhat inductive, and a smaller C across it can create a resonant circuit. Nice modern low ESR capacitors make for higher Q resonators and makes a whole lot more trouble than just the big C on its own.

Fashion sez you gotta do things that way and contesting an audiophile theory is like taking on the mafia. They all round on you, seeing any trace of substantial facts as a direct challenge to their entire universe.

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 2019, 8:05 pm   #13
kalee20
Dekatron
 
Join Date: Feb 2007
Location: Lynton, N. Devon, UK.
Posts: 7,060
Default Re: Valve preamp.

Yep. Big electrolytics of the 1950's would likely have benefited from a paralleled capacitor and it would have been good practice to at least consider the auxiliary component. Now that the same value is obtainable in a much smaller case size, and moreover measures are taken to reduce ESL (we have the switch-mode power supply industry to thank for this!), it is no longer necessary and could indeed be detrimental.

Methinks the designer has been slavishly following best practices of sixty years ago, without understanding why, in which case there's probably a lot more that he doesn't understand.

The constant-current load is, of course, an affront to the purists, but if there was a justifiable reason for including it, then it could be forgiven. As it happens, there isn't.
kalee20 is offline  
Old 9th Jun 2019, 12:47 am   #14
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Valve preamp.

Paralleling a larger capacitor with a smaller one is part of accepted audio practice alas. Even dafter is putting a smaller film cap across a larger film cap.

One way around potential resonance between the L of one and C of the other is a patent by Audio Research. They put an inductor in parallel with a resistor, and both in series with the electrolytic. A film cap is put in parallel with all that.

The idea is to deliberately decouple the electrolytic from the film cap. The resistor is in parallel with the inductor to damp any associated resonance.

Values from the patent are C(electro)=100uF, L=5uH, R=100ohms and 2uF film cap

Patent number US5,036,292 filed in 1990.

I attach the PSU section of the LS22 preamp which shows this used in practice.

Craig
Attached Thumbnails
Click image for larger version

Name:	ARC_LS22_schematic3.jpg
Views:	224
Size:	102.3 KB
ID:	184732  
Craig Sawyers is offline  
Old 9th Jun 2019, 7:53 am   #15
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Valve preamp.

It should have expired by now. Last time I looked US patents, with paid for renewals lasted for 17 years.

However... power supply distribution with paralleled reactive components will inevitably have resonances. The added inductors in that Audio Research circuit posted by Craig will add even more resonance modes. The trick lies in controlling the Q of each resonance so that it is not troublesome.

ESR and ESL of components is important so a schematic giving just capacitor, resistor and inductor marked values is incomplete, you need to get parts with the stray values the circuit was designed around.

Designers are often lazy about decoupling and just sprinkle capacitors around without thinking. New parts with extra-low ESRs create unexpected problems. In the past, the old ESR values did useful damping that few people were aware of.

People are often blinkered. It's an audio amplifier, so the automatic assumption is to think of decoupling at audio frequencies when the question they should be asking themselves is "What frequency range do I want it to be stable over?" and the answer is that you don't want it to be unstable at ANY frequency. So good design is to manage resonance Qs until you get to a high enough frequency that circuit and component losses can be trusted to spoil higher resonances.

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 9th Jun 2019, 8:39 am   #16
Diabolical Artificer
Dekatron
 
Diabolical Artificer's Avatar
 
Join Date: Jul 2009
Location: Sleaford, Lincs. UK.
Posts: 7,637
Default Re: Valve preamp.

Regarding paralleled C's in Pat Turner's circuits: from what I've read on his website he knows what he's doing and isn't an audiophool, and is/was unlikely to include bits of a circuit cos it's fashionable . He goes into pretty thorough analysis of circuits, tests them and went to great length's to knock up test gear.

I can't speak in any depth about the finer points of paralled caps and stability but PT has given me some good advice when building my big monoblocks, making them more stable, further his amp design's have a lot of thought put in, to make them stable not just over the audio frequency range.

Leaving aside the advisability of paralleling a big electrolytic with a smaller value one, which if any of the preamp design's in post #4 are any good/worth building?

Andy.
__________________
Curiosity hasn't killed this cat...so far.
Diabolical Artificer is online now  
Old 9th Jun 2019, 11:44 am   #17
PJL
Dekatron
 
Join Date: Sep 2005
Location: Seaford, East Sussex, UK.
Posts: 5,997
Default Re: Valve preamp.

Seems to be jumping in at the deep end. Here is my thread on a valve phono stage and there are plenty of simpler designs in there https://www.vintage-radio.net/forum/...d.php?t=102751.

I found hiss to be a problem with a pentode (EF86) phono stage so why not start with a simple 2-stage triode design with feedback? You can then add constant current sources or whatever and see if there is any improvement.

A design feature of most phono stages is plenty of LF roll-off to remove the distortion caused by warps in the record. The commercial IC design had a roll off of -18dB/octave.
PJL is offline  
Old 9th Jun 2019, 2:56 pm   #18
Craig Sawyers
Dekatron
 
Craig Sawyers's Avatar
 
Join Date: Nov 2010
Location: Oxford, UK.
Posts: 4,941
Default Re: Valve preamp.

Here's some modelling of Audio Research's capacitor with inductor etc.

Of course David is right. The best result is for just the electrolytic alone. The attached shows the comparison. I have chosen the values that Audio Research used in the LS22 - 6,600uF (3 x 2,200 in parallel). The vertical axis is the same in all three cases for direct comparison.

Craig
Attached Files
File Type: pdf Capacitor modelling.pdf (97.1 KB, 110 views)
Craig Sawyers is offline  
Old 9th Jun 2019, 5:50 pm   #19
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,800
Default Re: Valve preamp.

They say that sausages and laws are the two things that you should never see how they're made. Analysing decoupling networks could be added to that list. The old traditional ones and the new-broom-sweeps-cleaner ones all turn out scary.

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 9th Jun 2019, 9:27 pm   #20
Synchrodyne
Nonode
 
Join Date: Jan 2009
Location: Papamoa Beach, Bay of Plenty, New Zealand
Posts: 2,943
Default Re: Valve preamp.

Regarding the triode vs pentode issue, it seems to me that when it comes to RIAA stages, there is sometimes conflation between the characteristics of the valves themselves and the circuits in which they were used.

Typically the EF86 was used in a series feedback circuit. For matching magnetic cartridges, the input arm resistor was usually 47k, or 68k back in the day, and this, not the valve, determined where the noise floor was. Equipment maker specifications sometimes reflected this, and a typical example read “Noise: -80 dB, or where applicable, the equivalent noise of the pick-up load impedance at the input”.

The same constraint would apply if the EF86 were replaced by a single triode, or by a cascode double triode. But the EF86 was sometimes blamed for the noise shortcomings of the shunt feedback circuit.

And of course the shunt feedback constraint sustained in the solid-state era. H.P. Walker did the numbers c.1971, in respect of then available NPN transistors (BC109 & co.) and came up with theoretical best SNRs of 58.5 dB with respect to 2 mV for the shunt feedback case and 72 dB for the series feedback case, with a 50k input resistor. (Self later showed that the NE5534 op-amp in a series feedback circuit produced 84 dB relative to 5 mV, which is 76 dB relative to 2 mV)

That said, back in the day the EF86 was viewed as satisfactory for input sensitivities to around 4 or 5 mV, which was all that was needed in the 1950s, but circuit considerations aside, valve noise itself was seen as probably not low enough for sensitivities around the 2 mV level. And clearly the shunt feedback circuit input stage was inappropriate at this level; disposing of this was the primary requirement when approaching that 2 mV goal.

One way, advocated by Bailey (although I don’t think that he originated it) and used by Radford in its SC2 was to use a flat triode buffer stage at the input followed by another triode stage with shunt feedback RIAA equalization. The first triode stage operated without feedback. (If it had its own feedback loop, that would be of the shunt type, so would reintroduce the problem whose avoidance was the prime objective. Degeneration by omission of the cathode bypass capacitor would create a potential hum problem that would probably require DC heating to avoid.) Given reasonable gain in the first stage, a pentode was probably workable for the second stage, but as double triodes were the norm for consumer audio valves, using such was logical.

If DC heating were acceptable – and a minority of equipment makers thought that it was – then another approach was to use series feedback around a double triode – if you like it was the valve analogue of the Dinsdale two-transistor circuit. With the feedback loop returning to the first stage cathode, this was necessarily above signal ground and so susceptible to hum pick-up with AC heating. (Radford famously used a hybrid version in its SC22, with a BC107 first stage and half an ECC83 second stage, thus avoiding the need for a DC-heated first stage.) Possibly a triode-strapped EF86 for the first stage would have lower hum pick-up than an ECC83, so might have been workable with AC heating. (The Brimar 13D7 double triode apparently got close to the EF86 when it came to hum transfer levels.)

Where the job is simply to match and equalize MM cartridges, then applying Occam’s Razor would point to using an NE5534 or similar, such as in the circuits by Self and Baxandall. Fit and forget, then stop worrying. Anything else would be choosing the long route or the back road simply because you wanted to go that way – the scenery is better or the drive more challenging and so on.

By the way, something I haven’t seen – although it might well exist – is a scientific/engineering rationale in support of passive RIAA equalization. On its face it seems like another exercise in doing it the hard way, when by using a series NFB loop one neatly disposes of the initial gain and equalization requirements in one fell swoop.


Cheers,
Synchrodyne is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 8:05 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.