Next Single Ended build

[Paul Barker]

the problem is she doesn't spend money on herself.

[Mike H]

Oh, piddle

[Paul Barker]

Some progress has been made, no wiring underneath yet.



The R core output transformers are slung underneath so there isn't a large empty space as it seems from the top. There space between these transformers where I should be able to mount a filament transformer, which I'll wind to the exact spec of all the valves.

The main HT of 300v will go in through the RF socket shown mounted.

But now that I have made plans for two negative supplies for level shifting I shall have to make two new places for two different levels to arrive at the amplifier. All the DC levels will have to be created off board.

The first power supplies will be basically the last supply I used with a small amount of shunt regulation at the end to pull it down to 300v.

The two negative supplies: one is -500v so that will be voltage multiplier from 230v then shunt regulated with control to tune the level exactly.

The -200v will be normal rectifier with same arrangement.

So I have set myself a hard task in power supply demands.

But I bought the Tant's took the punishment, so I better use them.

[Paul Barker]

I do need to commend Anth, because the metal plates I sent him to make boxes for appear neither to be square nore exact corner angles. So he has made four corners of slight angle variation that fit snug. I only discovered this when I realised there is only one way the metal plate fits snug. the other three ways don't fit.

Anth never contacted me and even made mention of this, he just got on and made a perfect fit. Neither did he mention it when he'd sent the finished job. He is a consumate professional.

[cressy]

well i didnt think it worth mentioning, the plates were only out by a little so i just had to fettle the joints a little bit to get it to fit. it was less work to do that than it was to start filing the plates so they were bang square

[steve s]

Good job paul... Your PX4's seeing the light of day will be a suprise for them

Go carefully....

[Paul Barker]

Thanks both.

Steve, I woke up this morning head buzzing with all the fears.

I've gone and done it again, designed two stages of direct coupling to run together and depend on each other.

One very immediate decision is, this amp cannot be used for valve tasting. Once set up, it is to be left alone.

[Cressy Snr]

Thanks both.
I woke up this morning head buzzing with all the fears.

I've gone and done it again, designed two stages of direct coupling to run together and depend on each other.

I know just how you feel Paul.

Why do we do it?

[Paul Barker]

Mike H I hope you are following this thread.

I just put an order in for two of your 50va transformer kits. Is it possible to do a bank transfer email me the bank sort code and account number?

Anyone reading this is a no brainer.

I need windings of:

2 times 2.5v at 3.5 amp
2 times 6.3v at 0.3 amp
2 times 4v at 1 amp

Easy low number of turns Mike does a 50va kit where you just have to turn your secondaries, so few volts it's easy by hand. No brainer.

I only need one of these transformer kits but to feed the hoarder inside me I am ordering two

[Paul Barker]

Procrastinated too long.

Now changed my mind, it won't be direct coupled it will be Sukumarised. therefore the 27 is out the 10 is in.

Back to the patiently waiting for parts to turn up, and just hope I don't end up forgetting I have some Tantallum resistors and let them get lost in the medley of other electronic parts never to be found again.

Shouldn't have bought the posh parts before building it cheap first.

Never mind.

[Paul Barker]

OK so why did I bottle out?

I had it mostly planned, and pulled together some parts.

For the B+ of 300v I have the breadboard power supply from the last project which is about 330v if I remember correctly. A small amount of shunt current at the end would pull that down and improve regulation, so the main B+ was a done deal.

Looking at the sketch of the circuit below on the side of my philosophy the design ticked the boxes.



Here we see the only HP is the OPT and the LP is the series resistor with the miller of the second 27, which is over 50khz (thank you to Nick G for pointing this error in my original design out.)

I already bought the expensive 2 watt Tantallums 90k four of them. I bought some Tantallum grid stoppers.

The 50% loss of gain by the Tantallum level shifter was OK I had enough overall gain.

Where is the compromise in this design? There isn't one if you pay as much attention to the B- supplies as you do the B+ supplies.

The B+ is full wave valve rectified (potatoe) choke input Black Gate passive FFSR. So the B- has to be samewise but upside down.

So I found a 625-0-625 100mA transformer. Another potatoe no problem. Another suitable choke no problem. another passive FFSR no problem, even more Black Gate (now need two in series, might just have enough will have to search far and wide though.) Easy enough to make a fantastic active shunt reg at the end of the -530v and then devide down from that with VR for the first stage -200v.

So it could have been achieved. It probably would have sounded great. But it will have to go back into the recesses of the mind. One I can't devote time to right now in my life. As Nick said "is it worth it?" Not right now, no.

So what am I going to do instead.

Similar objective, but without the B-'s



The LP filters shouldn't be too bad as these are bifilar transformers just the cheap Hammond 126C.

HP filtering is now an issue as there are now 3 successive ones (the three transformers) which will definitely compromise bass performance.

We'll see how it sounds first shall we?

If LF needs help I can DC one of the stages. (pentode mu out as above perhaps with VR level shifting to maintain the gain, or use my Tantallum's which would probably sound better than the VR but I loose gain.)

What batteries are going in for the fixed bias? on all three stages? Will they affect sound? Probably. We'll find out won't we. will it sound worse than a bypassed cathode resistor? Might do, might not. That's why we try things, to learn the answers for ourselves.

My hopes and aspirations for this design are a wonderful organic sound. I am not necessarily after killer bass. My speakers wouldn't no what to do with it anyway.

No my objective is the best girl and guitar sound I can achieve.

I can't predict that it will be achieved. Hidden inside this compromised attempt are some HF peaks, I just hope that they will stay clear enough of my hearing not to affect my enjoyment. If you have the hearing of a batt tuff doodoo, it's not for you. But these inherent peaks on the bifilar transformers doubled up by the use of two may destroy my goal. Or they may not.

My bass phase is definitely pants. This might not matter to girl and guitar. But it might.

If the whole thing is pants and I don't even get gilr and guitar where I like her, I'll try something else. Nobody died did they?

Where are we going with dht heating on this design?

Might try AC on the two 10 stages and hope they self cancel with adjustment i.e. twiddle one until other cancels the pair. then use Andrews modules on the PX4. Might work, or else I have to order more from Andrew or use an idht stage one.

You might say "hay Paul you missed a trick use one of the sexy high transconductance valves instead of the two 10's, with your 126C and it's an easy solution done deal you fool"

Well that is the direction of Lemmings, I never follow Lemmings.The 10 has a sound, it is a sound I aspire to. People that use the one valve before the output valve approach cannot use the 10 or the 01a or any other known great sounding girl and guitar valves in a single stage. Even if they put this one valve in their power amp, they are kidding themselves, to get the gain required for full output they need an active preamp, so they have my amp anyways, but put it into two separate boxes with interconnects required.

The high transconductance route saves you a cap or saves you a transformer saves you a stage. But if there is a better sounding valve (like I propose to you the 10 to name one of many) two of that valve with correct handling of the coupling is hopefully going to sound better than your single stage of a good alrounder valve. The allrounder high transconductance valve will not (I protest) get the hair of your neck standing out straight.

The hair standing out is what I am aiming for. I might miss it, but I am shooting at it.

[Paul Barker]

All the time I am waiting for components gives my fertile mind time to plan.

So let's just look at the bass response and the phase angle which is also important for fidelity.

The proposed interstage transformers on order anyway have a primary inductance of 106 Henry and 375 ohm dc resistance.

The 10 internal impedance is probably 5k5 at the operating point.

So we are working with 5875 ohms. For fidelity Frits says we need the bass response down 0.5db which puts the phase at -15°

This point is 24.4 hz that would be OK on it's own for my loudspeakers. But when there are two stages the problem is additive. So instead of -0.5db and -15° we are -1db and -30° at 24.4hz.

So we are now pushing on the borders of phase affected bass at the kind of level full range speakers operate. We had better not have any further phase and attenuation effects.

Sadly we do. we have the output transformer. Just working with what I remember about it the OPT is 4k5 dc resistance 120 ohms primary inductance 15h. We shall use a guestimate for the PX4 internal resistance of 1kohm. You may think it might be less, but let's work with worst case scenario.

So with that lot we have -0.5db and -15° phase angle at 9.5hz. This is quite good, the OPT stand alone is good. But it's effectiveness is downgraded by the two stages of interstage.

We can make things even worse if we use autobias and don't get the right bypass cap.

So the objective of Sukumarising the amp has these compromises inherent in it. We are not expecting huge compromises at 20khz as the interstages are bifilar wound, we would expect a resonant peak where leakage inductance and the combined capacitances resonate, but we don't have figures for either. leakage inductance should be very low, but capacitance quite high. In the end tests will discover the facts about the capacitance of the bifilar arrangement. I expect the resonance to be far away from the audible frequencies.

But as it stands the plan is compromised at bass. There are two things we can do about this. We can use just one interstage and another method for the other. Or we could choose an alternative valve.

I decided to run the figures for 1/2 a 5687 per stage. These transformers suddenly look a lot better for the 5687. At my operating point the 5687 should be 2kohm so our -0.5db is at 9.8hz. Now we are back in the running for two transformer stages we maintain what Frits calls "fidelity" at our coveted 20hz.

To summarise, the Hammond 126C is marginal for loading the 10 in two stages (one stage it is acceptable but not astounding) but fantastic for loading 1/2 of a 5687.

That leaves us with the decision whether to use both and go with 5687 or to split the designs and use 10's but just one IT coupled. As 5687's sound pretty good I may aswell start with those and see if the proposed hair stands on end. One advantage with the 5687 is that it makes the design more accessible, though they are not cheap they are not has hard to procure as 10's, and certainly lower budget. The two 5687 stages can be battery biased. The output stage might end up standard 1k resistor biased bypassed. The cap has crept in.

[slowmotion]

Have you tried that bias arrangement before?

[Paul Barker]

[Paul Barker]

Have you tried that bias arrangement before?

Fixed bias with batteries? Yes. Are you asking because you want to know my experience or because you have some experience you want to share? Either way, speak up.