Parallel single ended valve amps, any experience?

[kalee20]

Originally Posted by Radio Wrangler

There also seems to be a 'wee discrepancy' between that peak current limit andthe power levels they're talking about, assuming class-A and something like 350v HT.

Thinking about it...

130mA standing current, Class A... let's assume a decent pentode or beam tetrode output valve.

Then the maximum downward current swing will be 130mA (and the upward swing the same, if you want undistorted sinewave).

So with 130mA peak swing, the RMS signal current will be 92mA.

If the transformer is set to provide 6kΩ load, then the voltage swing is 92mA x 6kΩ = 552V (rms) or 781V peak... we'll need HT at least equal to this, with 'ideal' valves to avoid clipping on peaks. (In practice, to allow for 20V lost across cathode resistor, and 30V cathode-anode bottoming voltage, we'd want a bit more, say 830V...)

The audio power out will be 50.7W (!) and the HT demand 101W (130mA at 781V.

Somehow I can't accept that anyone would make a transformer for this rather extreme voltage! So let's recalculate at 2kΩ anode load impedance.

Voltage swing now is 184V rms or 260V peak... we'll want HT to be at least that with 'ideal' valve.

Audio power output will be 16.9W and the HT demand 33.9W (130mA at 261V).

Neither extreme squares-up with the 25W output claimed, which for Class A would require minimum 50W power draw... 130mA implies 385V minimum.

I'm really dubious about these transformers... the 2k - 6k primary load, how can the effective load be varied? (A multi-tapped winding tends to compromise on effective coupling, because of the dead space occupied by unused sections of winding).

They could be OK, but the quality and self-consistency of the data doesn't exactly give confidence...

[Radio Wrangler]

That transformer spec says max current, not max standing current. So I took it literally and took 350v HT to be generous. Wind and tide behind and no losses didn't reach 12W. I allowed the anode to go down to 50v Anode dissipation a bit over 50W as well.

Quick wet-finger analysis (digit-al design?)

David

[kalee20]

Originally Posted by Radio Wrangler

That transformer spec says max current, not max standing current. So I took it literally...

I'd hope the 130-160mA was the standing current! But then, I'm biased because that's what I'd want in a spec.

The fact that we both can interpret the spec in different ways, justifiably, admits to a badly-compiled spec sheet.

And if the spec sheet has been muddily written, could it be a forlorn hope that the transformer itself has been designed more rigorously?

I'd certainly suggest to Gabriel to keep his chequebook in his pocket for now...

[Radio Wrangler]

It's a matter of risk. If things went pear-shaped m'learned friends would decide entirely on the literal statements of the spec sheet just what was being promised. I hope your view is right for the sake of people buying the things. I think it's likely that something got lost in translation.

David

[Gabe001]

I think there are "lost in transtation" issues as well, and having read quite a few descriptions in chinglish, Raa is actually Ra, max current I suspect is maximum standing current (not peak) and the impedance range is given because they can customise to any primary value from 2k to 5k (not multitap), and the current range is probably because they have 18w,20w and 25w versions which all use this same description. I only need it for 10w, so even if it doesn't stand up to the quotes figures, it will not be working anywhere near maximum.

Full link here (for the full chinglish experience)
https://a.aliexpress.com/_EQ6ABt1

If the transformer is not up to the job, plan b is a parafeed.

[Valvepower]

Hello,

Ah, getting parts from China…, I have to be honest, and speaking from experience, it can be tad hit and miss… but you can find some real gems!

The cultural differences come in to play with how some parts are specified and presented, so there is an element of reading between the lines – so to speak.

You can do your best as regards interrogating the specification, but if it’s a sensible price and it fits the bill, you just have to bite the bullet and get the part in question to eventuate it.

As you said previously, you’ll under run them…, which is a good plan.

If you’re looking at a standing Ia of 100mA, it’ll be interesting to see what valve(s) you’re looking to use.

Regards
Terry

[Gabe001]

2 parallel el34s in triode mode at around 450v plate and screen per channel, if I can, and all works out to plan. May be a bit less than 100mA in practice but as a ballpark figure, close enough. Should get me 5-6 w per valve depending on transformer losses - each valve will see a 5k load so the distortion and damping will be half decent, even before feedback. Driver and feedback amount still to be decided. I have a few ideas but they're just ideas for now. I'll put them in Lt spice and have a play.

Note: wimpy el34s need not apply. JJs maximum screen voltage is 450v, Mullard 500v. Chinese, who knows!

[ian_rodger]

Tried to follow link to these Chinese transformers and got the message below. Just for info.

This site has been reported as unsafe
Hosted by www.aliexpress.com
Microsoft recommends you don't continue to this site. It has been reported to Microsoft for containing harmful programs that may try to steal personal or financial information.

Go back
More information

[kalee20]

Originally Posted by Gabe001

Full link here (for the full chinglish experience)
https://a.aliexpress.com/_EQ6ABt1

I had a look! It does dispel some of the pessimism I expressed in previous posts.

There's just five leadouts, which I guess are primary (anode and HT+) and secondary (0 - 4 - 8Ω)

Basically, the different primary load impedances are determined as to what you order - that's good, the turns/wire gauge is chosen accordingly, and there's no 'dead space' as would be the case with a multi-tapped winding.

The secondary will have some wasted space, if you use the 4Ω tap, the remaining turns for 8Ω aren't used.

The photos of the bobbin seem to show evidence of interleaving - that's good, it improves coupling and reduces leakage inductance which would hurt top-end response.

Originally Posted by Gabe001

If the transformer is not up to the job, plan b is a parafeed.

That's a rather bad idea if it comes to that - unless you can remove the laminations and re-insert, interleaved. Then you'd get much higher primary inductance.

The overall inductance that the anode would see is the feed choke in parallel with the O/P transformer (assuming the coupling capacitor is large). The feed choke, which has to carry DC, can't be super-high as it needs its air-gap. But a 20H choke in parallel with a 200H transformer (18.2H effective) is going to have a better bass response than a 20H choke in parallel with a 20H transformer (10H effective).

Getting a suitable feed choke may not be trivial - the average filter choke deals with 100Hz ripple, and the harmonics taper off rapidly - there's very little above 1kHz. So techniques to reduce self-capacitance won't be employed (it puts cost up). But for full AF range, up to 20kHz, any self-capacitance will bypass the top-end frequencies unless the choke is wound to minimise self-capacitance.

[Gabe001]

Ian, AliExpress is fine, feel free to proceed to the site without worries.

[Richardgr]

The assumption at the start of this thread wa that paralleled tubes are necessary for volume.

If you use EL34 you also have the option of a SE OPT with a UL tap,and then you get plenty of power.

I built the Tubelab SSE with Toroidy output traffos, and there is more than enough volume.

[Gabe001]

Built one of those (in stereo) already Richard! Its gone to a family member. I liked it, which is why I want to build this one.

What's the tubelab SSE like and do you prefer the UL or triode mode? With feedback or without?

Kalee, thanks for that. I feel a bit more reassured. I suspect "golden segmentation cross" = interleaving

[Richardgr]

I am using the amp exclusively in UL mode since I have not felt the need to mess with the way it performs. However it is upside down on my kitchen table at the moment, so maybe I should get adventurous.

I think there is no FB, but I will confirm.

Ultimately sound comes down to OPTs, and the Toroidy transformers are a lot of bang for the buck.

[Gabe001]

There should be jumper settings for triode and feedback. Most people install switches so that they can change mode easily. You should have a play. What voltage are you running the output valve at?

[GrimJosef]

Originally Posted by Richardgr

The assumption at the start of this thread was that paralleled tubes are necessary for volume ...

Perhaps it depends on what we mean by 'paralleled' and by 'volume'.

I have a few Eimac 304TLs in store which, one day, I'd like to try in an SE amp. The 304TL's glass envelope contains four paralleled glowing-anode triodes, with a total anode dissipation of 300W (the filament(s) alone require 10V at 12.5A or 5V at 25A). There have been a few DIY builds online (mostly parafeed) which deliver 30W of audio without really breaking a sweat, unless you count what happens if you try to carry the power supply very far .

The attached picture shows me testing one of the valves at 1200V and 250mA with -66V bias. The Quad II in the background was a 100V line one which was a handy source of grid drive to check each valve's gain.

Cheers,

GJ

[Richardgr]

Yikes!

[Gabe001]

You need your own personal power station for that!

[Radio Wrangler]

Staying with Eimac (Eitel-McCulloch) there is their 3-500Z and withthose looks it ought to sound rather impressive. 500W anode dissipation, so you don't need to limit yourself to horn speakers though I guess it would be an amplifier for winter months. if the 3-500 isn't enough then there is the 8877/3CX1500 triode that roughly triples the power. I didn't need to say they're directly heated.

image from valve museum, where the datasheetcan be found.

David

[knobtwiddler]

Would it be too OT if we included an AF amp that uses beam tetrodes? This one has a 1KV PSU and caps rated at 3KV: http://www.technicalaudio.com/pdf/Fa...641_Manual.pdf

It's designed for cutterhead amps, but ought to be equally adept at driving speakers. I'd love to see the O/P transformer for it!

[Radio Wrangler]

4CX250Bs? Eimac again.

The $CX250B is well known in RF power amplifier circles. They were designed for non-linear Class-C operation, and intermodulation characteristics are not wonderful. People have used them in 'linear amplifiers' for SSB, but the linearity is only marginally acceptable. They are a high gain device (High perveance, too) and prone to UHF self-oscillation. The main defence is to use special valve sockets with integrated ceramic decoupling capacitors on to fingerstock connecting with the g2 ring. They can give peculiar behaviour like negative current at g2. So having g2 off to a transformer tapping is asking for trouble.

If asked to nominate a valv least likely to be used at audio and not allowed to name obvious non linear and switching wierdos, I'd have suggested 4CX250s.

Impressive work, if they've tamed those little devils at audio.

David