Leak TL12+ - component values for version with 8778 OPTX

[Lucien Nunes]

Perhaps some kind person with the data to hand could confirm for me, the correct component values for the version of the TL12+ with OPTX 8778 and mains TX 3920R. These differ from the version for which all hits on the WWW bring up. I can't find either my copy of Stephen Spicer's book or my old bench copy of the cct.

Specifically:
Feedback: I have both 180 and 200pf in parallel with 12k fitted.
V1 anode DC load: 47k fitted.
V1 anode comp network: 300pf in series with 4k7 fitted.

I have not put the component designation numbers because IIRC they differ between the circuit diagram I have and the correct one. Are these right? I believe all other components are the same for both versions, willing to be corrected.

Actually I would be happy to buy the service data from the download area if someone could confirm whether it includes this version.

TIA Lucien

[DangerMan]

I have a LEAK Stereo 20 installation/operation/maintenance document which shows both circuits.

To answer your questions in order:

Feedback was 12 kohm with 200 pf in parallel for both versions.
Anode load for V1 is indeed 47k for the later version.

But,
V1 compensation step network, is shown in this particular document as not present for either version, which is also true for my own later-version amplifier. (Gold painted)
The later-still versions (Grey painted) do have the network as you specify.

Plus,
Grid leak resistors for the EL84’s become 470k in revised version instead of earlier 1meg.

Finally,
Capacitor C5R which decouples V2b grid to ground changes in updated version to 0.1uF from 0.02uF.

Hope this helps

Pete

[Radio Wrangler]

The grid coupler capacitor increase is there to accompany the drop in grid leak resistor and keep the LF roll-off the same.

The driver/splitter stage will feel a heavier load from the grid leaks it is driving so some changes may have been needed back down the strip.

David

[Lucien Nunes]

@Dangerman that is helpful. I think some versions of the circuit diagram incorrectly show 20pF in the feedback path, which might be why I recall some scribble there on my lost copy of the circuit.

I have 0.1μF for the phase splitter ground reference, 470kΩ grid leaks and 0.25 (now 0.22) μF output coupling caps.

Just need to confirm the value of the compensation network cap, which I have as 300pF but the other version shows 200pF.

These amps have not been used for a few years and the 8/16 smoothers have given out in the meantime. While they are on the bench I want to undo some mods made around V1 when I first had them as a teenager. I lowered the gain but probably wasn't too careful with the stability. They look quite ingenious though - I see that you can have either stock or modified config according to how you wire (non-standardly) the octal plug with a couple of links. They have travelled far and served me well, and have many battle scars to show for it. But now I just want them back to original to be enjoyed at home again.

[Herald1360]

The TL12 with triode connected KT66s did have different feedback R and C values for different output TX tapping settings. Maybe some confusion here?

[DangerMan]

I have found my copy of Stephen Spicer’s book, so can tell you what values are shown in his copy of Leak’s diagram for TL12+.

Although he doesn’t confirm which variant it is, or transformer types, the circuit does show a 280-0-280 mains transformer which I think makes it a modified version. The diagram looks like a Leak drawing, and is reproduced with permission.

EL84 grid leaks are 470k, grid coupling capacitors 0.22uF, phase splitter cap 0.1uF.. all as expected.

Anode load for EF86 is shown as 82k and step network is 4.7k series 0.0002 uF (200pf).

Feedback is different though, 12k in parallel with 0.00002uF (yes, 20pf)!! This seems likely to be a mistake for 0.0002uF and made on the original...?

A bit of a can of worms when the devil is in the detail.

Pete

PS
If it helps, I’d be very happy to scan and email this diagram to you.

[Lucien Nunes]

Thanks for the offer but I think this is the circuit that I have printed from an online source. It is for an amp with a 280-0-280 mains TX, values are as per your comments and include the highly suspect 20pF feedback cap which does seem to be an error. I think we're looking for the 'other' diagram, to validate or condemn the 300pF step network cap when used with the 47k anode load.

My suspcicion is that I was playing about with this end of the circuit and the caps now fitted are RS mica types which might not be original, but the values might be. In the absence of a circuit diagram or parts list, perhaps someone might have an early amp they can flip over and check.

[Valvepower]

Hello Lucien,

I have a pair of TL12+'s with those transformers, which I restored a while back. I also have earlier TL12+'s with the first variant of the transformers as well.

If nowt else appears I'll have look this morning after doing a job for SWMBO and report back.

My late friend Steve, who worked for Leak in the 1960's, did pass on some verbal information regards variants etc., which helped with the 'odd-ball' stuff.

Regards
Terry

[Valvepower]

Here we go…

Pictures of my two Leak TL12+’s with 8778 output transformer in the attached zip file.

Using the component identifiers in Steve Spicer book:
R5 = 47K
R23 = 4K7
R12 = 12K
C15 = 300pF
C7 = 220pF (this may have been 200pF in the original, which I’ve replaced with the nearest preferred value of 220pF).
C1 = 1000pF

I figure when the UL tap was changed from 49% to 25% (3921 to 8778) the open loop gain and suchlike would have changed very slightly so the compensation network needed tweaking. Also, at this point there was less focus on the valve amplifiers, as the future was seen as transistorized, and the focus on the valve amplifiers was less stringent (for want of better words), so the ‘pass’ figures could have been relaxed very slightly, whereas in the early days of the TL12+ the pass figures were kept tight in pursuance of low distortion and power output etc., etc. This gets coved in Steve Spicers book.

I remember being told Leak improved the stability margins on the later valve amplifiers.

I have seen some ST20 amplifiers with no compensation network on the first stage anode resistor, whereas they were used on later ST20's, however they also used the 8778 output transformer.

It was good to look at the Steve Spicer book again as it does bring back some fond memories of my Late friend Steve

Anyway, hope this helps a bit.

Terry

[Lucien Nunes]

Smart work in those amps, Terry, and thanks for taking the trouble to look underneath them. Your component values match mine exactly, and yes I have an original 200pF Hunts polystyrene feedback cap in one amp (rather than 220) and 180pF mica in the other that might be a replacement.

I see you opted for a film cap for the V1 screen bypass, makes sense, for some reason I stuck with electrolytic at greater expense and no benefit. This afternoon I will change the dead green 8/16's and reform the main cans and see where we are. I just checked the Philips 040-series 50@40 cathode bypass caps that I put in during the 80s and they are so low in ESR they might as well be pieces of AC-only wire and spot-on in value. They were good caps, I used them in everything until my bag of 100 or so ran out!

[Valvepower]

Hello Lucien,

Thanks for that, I suppose I aim to get as a neat an assembly as Leak would have required when it was first made.

I use what I call commercial components, which are available from the likes of RS, Farnell, CPC and Digikey etc.

Using metal film MFR5 resistors may have me struck of the audiophile list, but ‘hey’ they sound OK to me

I tend to use film capacitors in the g2 position (if they are available at sensible price though) because of the high value of the HT to screen resistor – I suppose it a personal foible.

Strangely, I like to use just one component from the original, however it must test perfectly OK, but, if not, I don’t do it and use a new part. It’s a bit like a Luther replacing the ‘mouse’ found in violin when restoring it.

Finally, Steve told me stores of his Mum doing outwork for leak and wiring and stuffing tag boards on the dining room table!

Terry

[Valvepower]

Hello,

As a bit of therapy to divert one’s mind off the current problems, I tuned into the Johnnie Walker show 1972 special yesterday, turned up the volume on the workshop amp and put the two TL12+’s with 8778 o/p t/x on the bench, and run them up to see how they measured, as it had been a bit of time since I last measured and restored them.

I fitted them with two pairs of matched Billington gold EL84’s and the rest of the valves were from the toy box. As well as checking the power supply and valve voltages etc., I also note the voltage between the two anodes of the ECC81 phase splitter.

It’s sometimes claimed they can’t actually get the 0.1% THD at 12 Watts so I setup the test gear and did a THD and frequency response sweep, and I got as near as dam it 0.1% THD into 16Ω at 12watts, the frequency response at 1 Watt isn’t too shabby either with a -3dB point at 45KHz.

Graphs attached.

Terry

[Lucien Nunes]

I think we can agree they meet the spec for all practical purposes. Time permitting I will measure these as well. There should be a couple of decent sets of valves lurking around the tester, as I had some good pairs left over after making up quads for two S20s where I wanted matching appearance as well as characteristics. As these are likely to start clocking up some hours I am tempted to park the Mullard GZ34s and let them run with something cheaper.

The main electrolytics came up well yesterday. Although there is slight deformation of the rubber bungs they are obviously still well sealed as the leakage, ESR and capacitance are all good as new. The 8/16s were completely shot though, no discernable capacitation going on in the outer section of either of them.

[Richardgr]

I fitted serial silicon diodes on the anodes of my Quad tube rectifiers. It is a trivial change that give a bit of belt and braces protection to the rectifier, and makes fitting a cheaper replacement less risky.

[bikerhifinut]

I used a Chinese "Gold Dragon" (PM Components house brand) GZ34 in my venerable and much loved Stereo20 when the original Mullard failed. Given no trouble over the last 5 years and given that the TL12 plus will only place "half" the load on a GZ34 I'd expect a modern JJ or other Current production valve to last well. In "lesser" makes (Ahem) of this fairly conventional EF86/ECC8*/EL84 push pull amp you'd find an EZ81, which would have been well up to the biz. I believe this is what is specced for the 5-10.
Just my Pennorth Lucien, and I am four square with my old mate Terry on all the info supplied. We were both on the original yahoo Leak Group before it was closed down. Probably a few others inhabiting this forum.
Despite some of the "Knockers" on web groups and certain publications, I always felt that HJ Leak had achieved the best compromise of performance versus price in a competent and practical domestic audio design.
So much so that I based my Homebrew EL34 power amp heavily on the leak circuit which is now my main amplifier in the current stereo system.
All the best Lucien.

Andy

[Lucien Nunes]

@Richardgr makes a good point that the addition of 'safety diodes' is easy and cost effective, but I also agree that any GZ34 ought to be OK in a TL12+. It's certainly an easy life compared to some stage amps. I will use the existing Mullards for now as they test good, although I worry about the symmetry of my sound stage as one valve has a fatter bass than the other. At least, it has a fatter *base* than the other, perhaps that's different

I was going to run these up last night but forgot to bring any EF86s back from the warehouse. I try to avoid randomly borrowing valves from things so resisted the temptation to open up any other gear. I'll be back there tomorrow so it can wait. The heap of EL84s at the test bench were a little disappointing because there are some matched trios, ideal for those push-pull-push circuits. Four close pairs to choose from with emission and gm in the 'nearly new' range; Electro-Harmonix, Brimar, unknown Soviet and (cough) Bentley. I'll see what else is at the warehouse.

[Lucien Nunes]

I am pleased to say that these TL12+ amps are now back in service. I took advantage of some post-chemo downtime to amuse myself quietly tidying them up. When I got them as a teenager in the 1980s they were not in high demand and I did not see them so much as vintage items, more as good general purpose small valve hifi amps. They have seen a lot of life and use in at least ten different locations and I did not think it appropriate to try to diguise that history, manifest through scuffs and scratches.

When I first got them, I changed most passives except the main electrolytic cans, 8/16μF decouplers, HT smoother wirewound, and mica and polystyrene feedback and step network caps. All of these are still OK except the 8/16's, which are now replaced by Vishay (of Philips heritage) 10's and 22's. Remaining from that earlier refurb are the pink SRC polyester coupling caps and the 040-series electrolytics in the cathodes, all of which are good components in good condition. Gone are the rather puny 1/4W CF resistors I had fitted, replaced with 2W MF.

One chassis had already had its first stage valveholder replaced but I am not now happy with that replacement. It has crackled a couple of times and I intend to replace it again with a matching McMurdo.

Valve complement at the moment includes a two totally random EF86s, one Soviet, one Philips; I'll probably swap one or other out before doing measurements as they look a bit of an odd pair if nothing else. My stash of ECC81s was rather lacklustre with many examples lurking in the 'emergency' stock; valves that work but for which nothing more good can be said. I ended up with two excellent but different valves, one Brimar, one GE, and again for the sake of appearances I will probably fit two the same once I get a few more tested. Two pairs of EH EL84s measuring as-new are in there to stay, while the two Mullard GZ34s measuring nearly-new will do for now but will probably get retrieved for use in a more demanding role in the future.

Finally I have them on this neat little rack that allows good airflow and avoids them heating the shelf above too much. Driving the dining-room late-version Kefkit 3's. The big Pioneer SX is sulking now.

[Valvepower]

Hi,

Nice job. Being battle sacred is OK, in fact some of the best Leak amps I’ve heard have been battle sacred examples!

I use a pair of KEF KIT 3’s when I’m not using the Rogers LS3/6’s. My KEF KIT 3’s has the oversized cabinet, and the Falcon Acoustics KEF Concerto crossover network.

I find TL12+’s and ST20’s drives the KEF KIT 3’s OK, however if I need more valve power, I’ll breakout the TL25+’s.

Terry

Originally Posted by Lucien Nunes

I am pleased to say that these TL12+ amps are now back in service. I took advantage of some post-chemo downtime to amuse myself quietly tidying them up. When I got them as a teenager in the 1980s they were not in high demand and I did not see them so much as vintage items, more as good general purpose small valve hifi amps. They have seen a lot of life and use in at least ten different locations and I did not think it appropriate to try to diguise that history, manifest through scuffs and scratches.

When I first got them, I changed most passives except the main electrolytic cans, 8/16μF decouplers, HT smoother wirewound, and mica and polystyrene feedback and step network caps. All of these are still OK except the 8/16's, which are now replaced by Vishay (of Philips heritage) 10's and 22's. Remaining from that earlier refurb are the pink SRC polyester coupling caps and the 040-series electrolytics in the cathodes, all of which are good components in good condition. Gone are the rather puny 1/4W CF resistors I had fitted, replaced with 2W MF.

One chassis had already had its first stage valveholder replaced but I am not now happy with that replacement. It has crackled a couple of times and I intend to replace it again with a matching McMurdo.

Valve complement at the moment includes a two totally random EF86s, one Soviet, one Philips; I'll probably swap one or other out before doing measurements as they look a bit of an odd pair if nothing else. My stash of ECC81s was rather lacklustre with many examples lurking in the 'emergency' stock; valves that work but for which nothing more good can be said. I ended up with two excellent but different valves, one Brimar, one GE, and again for the sake of appearances I will probably fit two the same once I get a few more tested. Two pairs of EH EL84s measuring as-new are in there to stay, while the two Mullard GZ34s measuring nearly-new will do for now but will probably get retrieved for use in a more demanding role in the future.

Finally I have them on this neat little rack that allows good airflow and avoids them heating the shelf above too much. Driving the dining-room late-version Kefkit 3's. The big Pioneer SX is sulking now.