trobbins

Jim, did you look through the Silicon Chip references to add some insight to your faultfinding effort?

Perhaps google the transformer makers, and include transformer or inter-stage transformer or the label on the transformers. Emmco Electrovoice is another brand/range. References like RDH3 describe that form of transformer coupling. Apart from microphone and output transformers, the last coupling transformer still in common use in the 1940's was a phase-splitter transformer that fed a push-pull output stage. Audio magazines from the 1930-40's have articles - mainly related to radio application, but also PA, that would be relevant. Perhaps the largest on-line compendium of magazines is worldradiohistory, and even magazines like https://www.worldradiohistory.com/Australian-Radio-World.htm are worth searching through, or keyword searching (eg. https://www.worldradiohistory.com/hd2/IDX-AUSTRALIA/IDX/Archive-Australian-Radio-World-IDX/Australasian-Radio-World-Vol-08-No-04-1943-09-15-OCR-Page-0011.pdf#search="interstage transformer"). Many transformer references go quickly into theory and equivalent circuits, which may not be helpful, but hopefully a few practical examples of what turns ratios are needed for different applications may help. Schure has a book on transformers that may be a good start too - https://www.worldradiohistory.com/BOOKSHELF-ARH/Technology/Rider-Books/Transformers - Alexander Schure.pdf . Some transformers indicate a turns or impedance ratio - you will have to come to grips with what they mean and how to measure turns ratios and then convert to impedance ratios. Do you have a meter for inductance measurement?

They are all likely inter-stage coupling transformers from the days when that form of coupling was common. You could try and re-create very vintage preamp and power amp etc style equipment by searching through early on-line magazine articles for suitable designs, but even then you would need a good dose of experience to redesign for commonly used tubes. Are you keen to do some homework?

Same as the STAR SA-30 on hi-fi engine. I have restored my unit. Lots of small safety, reliability and operational issues to correct imho. The outputs were not unconditionally stable for default 13dB of feedback, and I had to use Patrick Turner's technique of adding some current feedback, which then allowed the first stage step-network to be removed and can provide +/-1dB to 20kHz with suitable tone setting.

Managed to iterate an inductor value of about 0.7uH to give a reasonably smooth frequency response out to nearly 200kHz, and is unconditionally stable for capacitance only loads. No compensation capacitor across the feedback resistor was needed for this. Hats off again to Patrick Turner who afaik pioneered this form of global negative current feedback for valve amps to make them unconditionally stable. The worst squarewave response I could elicit was with 47nF load causing about a 50% overshoot ring at 100kHz that was quickly damped in about 4-5 cycles. The gain-phase plot for that condition is also attached.

I've been able to check the stability performance of my 10-10 with its A&R 4005 output transformers and 15 ohm load. The input stage has been changed to a 12AX7 triode (compared to original EF86 pentode) with 39k feedback resistor providing 20dB feedback, rather than 60k for original cct. With 13dB feedback and no comp cap the output is stable but shows a dominant resonance at 90kHz, and the response can be tuned to flat out to 120kHz with 30pF comp. With 20dB feedback the output is not stable without a tuned compensation cap, and circa 60pF achieves a flat response out to circa 150kHz. So that is somewhat consistent with the original circuit's feedback network, and with Johannus's comments. With 20dB feedback, the amp is stable with no load, but any capacitance only loading causes unstable operation. So I'll add some current feedback, as per Patrick Turner's examples, as I've also confirmed that technique allows unconditional stability with capacitance only loads in a few other amps, and is quite simple to implement. Of note with the Mullard 10-10 circuit (and many others) is that it uses a long tail phase splitter, which adds another level of checking when choosing 6GW8 pairs - not only should the pentodes be reasonably matched, but also the triodes, which can make the valve selection process tougher.

Just an update that i have my 10-10 back on the bench, and have a new tool to measure gain-phase of the power stage with the A&R 4005's, so will be interesting to assess stability and performance with different loading, and reflect on Johannus's comments from 2014.

Andrew, do all six channels use the same circuit/parts/heatsink, but a common Q6303 thermal feedback signal ?

The amp is described as 2-channel, but then there are L, R, centre and surround channels identified as having their bias settings changed?? The L and R channel bias is increased ~40% (60 to 82mA). That may certainly increase idle junction temp, and change cross-over distortion, but no info given on either except for subjective testing - perhaps an indication of heatsink temp rise may be worthwhile, as each device is idle dissipating another 1.2W. Is Q6011 thermally coupled to Q6051/Q6061 ? Have you tried doing harmonic distortion testing on a shoestring budget using your soundcard and some audio software ?

Many pots have a very loose tolerance - ya get what ya pay for.

It would, although it needs practical consideration of where to fit the fuse, and what value/type of fuse to use - which may require some expertise to adequately do. Imho it would be worth using a light-bulb limiter to provide some confidence that any changes you make (including output stage valve swaps) haven't accidently gone wrong.

You may want to consider retrofitting a fuse in the power transformer secondary HT winding, to mitigate collateral damage from an accident or poor tube or arc-over. It would be sad to lose a PT or OPT, as I guess there are few around for replacements. But the fuse has to be sized appropriately which would require some effort to measure winding and choke resistances.

If you have a known 30k trimpot setting that is ok, and you measure that trimpot as 3k5 between the circuitry nodes then you could replace the 30k with a 10k or 5k trimpot to get more adjustment sensitivity. But if the trimpot connects to the circuitry such that the circuitry sees a resistance of 26k5 from the trimpot then you could replace the trimpot by say a 22k fixed resistor in series with a 10k trimpot depending on how dextrous you are with fitting parts. But you need to inspect a schematic and the pcb tracks to confirm how the trimpot is actually connected. As a failsafe measure, most trimpots have the wiper connected to one end of the trimpot, so that a poor wiper doesn't cause an open-circuit (but rather just a resistance equal to the trimpot resistance).

Hi Matt, I saw that recent ebay listing. It would be great if you could identify the feedback network deployed in your amp, along with A&R OT2644 basic details (as they are not in any publication I have) and even if you can measure the key detailed OPT parameters.

Updated mirror. https://atrad-audio.co.nz/turneraudio/www.turneraudio.com.au/index.html

Arrgh, just went searching through my info in more depth and can see the TA1376B has a rating plate on the side showing 6k6 PP impedance. Also it includes a feedback winding and a shield, which my OPT doesn't have, so that ends that possibility. Which then indicates the OPT may be a more commercial 5k6 PP UL with 4/16 ohm output windings.

I have a Trimax UL output transformer without a rating/name plate. There is very limited catalog and other details on Trimax output transformer offerings, and the only reference I can identify so far of a Trimax UL OPT of circa 10-20W rating is to the PMG Type 3 amplifier made by Trimax, so I'm hoping that someone has a type 3 amplifier (A54 or A54A version) and has made, or could make, some measurements on the TA1376B output transformer in that amp please. The unknown transformer I have is in an M53 style casing that is 4" tall - so likely the M53-5 case (as identified in the A54/A54A manuals) although the only Trimax catalog I have showing M53 case information stops at M53-4. The rating plate was affixed to the top surface, not as a side plate, which also aligns with the Trimax labelling used in the Type 3 amps. The transformer has turns ratios that align with a 4.2k PP primary to 3/12 ohm secondary, with UL taps at 50% turns. The A54 and A54A amps identify the TA1376B having 3/12 output impedance windings, but don't specify the PP or UL primary details. The typical EL34 datasheet info relates to the original Hafler/Kereos 43% turns ratio UL setting and use a 6k PP Ra-a and 470 ohm per cathode auto-bias. The A54/A54A uses about the same per cathode resistance. Moers shows measurement results that indicate insignificant distortion difference between 43% and 50% turns settings for UL with EL34. The simplest measurement I could suggest is the P-CT-P dc resistances, which for my OPT measure as 136+136 ohm (which could vary a bit between OPT samples and due to temperature). The same measured resistance for each primary side is a good indication of hi-fi like symmetric windings. The next simplest test would be for turns ratio, and would need any external speaker/load disconnected, and a test 50Hz low-voltage connected to one section of winding and voltage measurements made to the other winding sections, but that is somewhat onerous for anyone who hasn't done that test before. I can set up to do impedance testing to check for primary winding leakage inductance and capacitance and resonances, and a frequency response test, but that would take a bit of time to arrange (I did a batch of testing on Williamson related OPT's last year), but is another indicator of performance of this OPT.

There could be many reasons for some remnant hum, and chasing hum can be hit or miss when making changes. Most chasing and changes can get quite technical - so not something to suggest unless tools and experience are at hand. Even before starting any chase there needs to be an awareness of whether the amp originally had any such noticeable hum, and if it didn't then what has since changed.

Well worthwhile ensuring adequate access to cool ambient air for those output tubes. KT88's have a wider glass envelope than the 8045G, and there is noticeably less separation with KT88 or other similar tubes, making it more difficult to keep glass temperature sufficiently low. Luckily the photo in the initial post shows that KT88 anode wings point to free ambient, rather than to each other, and that the KT88 heater dissipates 6W less than the 8045G.

Patrick's website is presently mirrored on: http://naturetour.digicom.bg/index-2.html

Back-bias was rarely used after circa 1950 for larger power rated amps as solid-state diodes started to allow practical fixed bias generation. Also the effective cost and performance of a phase inverter tube and parts started to overcome the quite common driver transformer technique (and of course Ferguson made driver transformers!). A serendipitous gift of an amp can often build incentive towards a tangible goal comprising vintage parts. Just be mindful of the effort and tools needed to renovate such an amp - a learning curve may be needed, but at least you wouldn't be learning on an expensive vintage amp.

It's not that you can't use two 16 ohm speakers in series as a load between the 250 and 500 ohm taps. That secondary winding section represents 43 ohm for an assumed 3k8 PP, and primary side would still be happy with a 2k9 ohm PP load, given everything in valve land is +/- 50% tolerance give or take. If it had a 125 ohm tap like that Ferguson then options would be simpler as a 16 ohm speaker would be fine across the 125 to 250 ohm taps. Many PA amps can achieve a wide bandwidth and suit a nominal speaker with a bit of rework.

Yes the 60W Ferguson schematic is a good start as it uses back-bias. Are you able to work out the circuitry on your amp? It's not too difficult to use MS Paint to modify such a schematic to show exactly what your amp has. The transformers/chokes look like they are PYE in-house, but yes the OP-18 primary 3k8 ohm PP would be appropriate. It looks like the choke has practical markings of 20H, ?50mA. The other transformers appear to have good terminal markings.

Perhaps check the link addresses, as those links don't work for me.

The amp appears to date from the late 1940's, based on the Erie ceramic tube resistors used for the 807 screen stoppers. It likely uses back-bias for the 807's. Although the Williamson had just come out by then, most amps had no feedback and had to seriously roll off the high frequencies (ie. the 807 plate capacitors) as audio material often had raucous treble content.