6550 SE Amplifier

[Nick]

They are your valves, so up to you.

I manually turn heater supplies up then b+. Two outputs on the variac, all heaters first, then b+.

I always assume a unexpected short term power outage could happen, so I either make sure that the startup control is automatic, or not needed.

The chokes arent instead of the 80k it is added.

Yes, I know that, but it was the startup condition I was referring to.

[Paul Barker]

The Quads seem as efficient as the kEF’s. This 8/9 watt amp drives them good.

[Max N]

They are your valves, so up to you.

I manually turn heater supplies up then b+. Two outputs on the variac, all heaters first, then b+.

I always assume a unexpected short term power outage could happen, so I either make sure that the startup control is automatic, or not needed.

The chokes arent instead of the 80k it is added.

Yes, I know that, but it was the startup condition I was referring to.

Would a 350V zener from anode to cathode work as a safeguard?

[Nick]

Would need at least a 5W zener, but yes I guess so. But remember Max, adding some solid state may cause all manner of demons to fly out . A potential divider before the first stage would do the job, a 88 doesn't need a 80k load resistor, especially if its got a choke in there as well. Or just leave it as it is and at worst replace a 88 after every power cut

[Paul Barker]

I see, so if there is an outage, the power supply would come straight on without enough current demand for critical inductance and momentarily reach 1.414. So I need a slow start for the valve diodes.. there was a slaw start idea in a Glass Audio letter. I think it was to pass the ground connection through a slow warm up valve cathode follower. Maybe 6AS7. Or there is always that delay valve dls10. Pretty sure I saw one while I was looking for something else.

I’ll work on it.

The reason I went for the maximum anode load was because I wanted to use cathode feedback and needed to maximise load for sufficient gain. I have sufficient gain.

[Nick]

Even without the 1.414/0.9 = 1.57 excess voltage, I was thinking more about the cold start condition of the ECC88 that will see the entire B+ on its anode until it starts to pass current.

[Paul Barker]

Yes, and probably the 6550’s are slow. Its just I’d not thought of the power out condition, but I suppose it is going to occur sooner or later.

Call me a Luddite but Ive just ordered a dls10, even though I know I spotted one lately. At £17 on ebay delivered, not a huge cost, meanwhile I’ll hope there is no power out. I’ll start the rectifier supply through it.

[Paul Barker]

Thanks for pointing out the risk.

Now that this amp is bedded down on the top floor I’ll have to build something for ground floor.

[Paul Barker]

[Paul Barker]

The good news is non of the output valves are sounding any different but a couple GE’s are 5v down in bias. The 4 Sovtec’ all ST62 are 5v down including the fourth which was never used. The GE’s sound better and drawer more current. But the Sovtec’s are all the same as the new one of the four. So no harm done to Sovtec’s. Nobody would hear any difference in any of the 10 valves.

The dls10 arrived. Just have to mount it in the power supply and find a transformer for it. Meanwhile turning B+ after 30 seconds.

Still got loads of these GE output valves not tried. Have yet to try the ST62 Tungsol British made. Another day. Might get some anode caps and try the TT’s.

I dont hear any 2nd so maybe the ecc88 is cancelling the worst of it? Answeers on a postcard.

[Nick]

I dont hear any 2nd so maybe the ecc88 is cancelling the worst of it? Answeers on a postcard.

I find in the absence of test kit, the clarity (or lack of) of the voice in this is a good measure.

[Mike H]

I always assume a unexpected short term power outage could happen, so I either make sure that the startup control is automatic, or not needed.

I had exactly one of these yesterday - Western Power have been 'fiddling about' with the underground network over the road for weeks, but this was the first outage, power was off most of the morning, back on 2 hours earlier than advised, OK fab. Then went off unexpectantly (no warning) for 5 seconds in the evening. Too short a time to get up and turn the system off in case there would be 'consequences'. There weren't any, fortunately. (But years ago I had a cheap DVD player go toes up due to exact same kind of few seconds outage. Looking inside afterwards, the very simple (basically a power oscillator) SMPS' single power transistor had blown. Or the winding, can't remember which.

I condidered a possible repair, but TBH it was too cheap to be worth it, just buy another one.

[Paul Barker]

I found the electro harmonix EL34’s sound much better. But thats a bonus because a single New Boxed GE 6550 /KT88 £195 on Ebay. So Ive got 8 and 6 used but work like new 4 sovtec and two Tungsol made in Britain. so I can sell those, since EL34 much better sounding.

[Cressy Snr]

[Paul Barker]

At the moment single ended one valve has a 4k6 load, but I drew the above to see what PSE looks like but the opt’s are 2k ohm. This graph represents 4k because two EL34’s achieve double the power. So power out becomes 16 watts. Don’t ask me where the distortion ends up, but I expect that when you check distortion at the same power out, 8 watts it would be less.

Interesting that the power out as drawn with one valve pretty well agrees it shows to use a 3k ohm load 348 Va 6 W 8% 2nd harmonic data in the Vade Mecum. Its a little anoying a lot of charts and data books dont specify what the Anode impedance is when triode connected. But the fact Vade Mecum suggests a 3k load, triode connected which implies a anode impedance between 600 to 800 ohm on the same lines of 2a3 to PX4. For the px4 a 600 ohm cathode resistor biases it. Though my choice of cathode resistor (which is per valve in the case of pse) is similar, right now Im using the 618 ohm. But the maths for the chart shown is 677 ohm precicely. The resistors are 680, so that will do and the 6k8 parallel is surplus to requirements now though not snipped off yet, as it is dual purpose to run with 6550 or EL 34, but I love EL34 more. 6550 is clinical, EL34 is emotional . I have shifted the operating point to the maximum voltage and set a dissipation limit of 27 watts. To allow for a little of the 8 watts permitted in the g2 to the 25 watts allowed from Anode. I could measure it in reality. But havent, by checking the current across the resistor connecting g2 to Anode. But haven’t. I’d be happy with ensuring the total combined current stays below 28 watts.

The power supply is capable of the current demand of pse. I think the 2k C cores should cope with the 116 mA. Hope so anyway. Time will tell.

I think I’ll keep this single ended version as built, and make another signal section for the pse, as the 2k transformers are existing that I made for the 6as7, about twice the weight as the r cores on single core like the r cores and Lundahls. I designed more primary inductance to make that planned amp at the time excellent at bass. But also interleaved 8 layers for good bandwidth. Made about 20 years ago.

There is no hurry 6 watts at home is OK even on the Q57’s but I do have to have the volume dial turned up to max. Which makes me consider putting a switch to switch out the 100k log Panasonic pot for grid chokes for such times when I’m alone.

Further experimental ideas. I found 6 hammond 150h for 8mA anode load chokes. Whilst I am totally happy with the sound of the amp, I may add three per channel to the resistor stack anode load in this amp. Just in case it helps the organic sound the gain and the bass. Probably in the present version. The beauty of inductors is as they increase in number series connected the capacitance reduces to half and inductance doubles per doubling of chokes.Nice!