Please check this PSU

[cbutterworth]

After vacillating between new amp for office and tweaking my at home Aikido, I have opted for the latter, as it'll be less expensive.

At present, my Aikido (running 4 X 6SN7) has a current draw of around 20mA, boosted to 34mA with bleeder resistors across the final cap. B+ under load is around 290V.

My B+ transformer has a slight buzz. It is one of the Hammond 260 series (125-0-125) with 115V primaries but actually getting 120-124 VAC.

So replacement with a 300 series Hammond would allow for the primaries to accept 120V. And why not change out my SS bridge for tube rectified at the same time?

Using a Hammond 275-0-275 150mA into a 5V4GA; 1.5uF cap; 10H choke; 60uF cap; 10H choke; 100uF cap, I get pretty good voltage and low ripple. But what bad things should I be looking for?

Please anyone, check this out, I can afford the new iron, but want to get the correct value first and I can get a 100uF cap at my local surpls place.

The .psu file has been renamed with .txt. I hope it works in PSUD.

Thanks,
Charlie

[Sheldon]

Why so much current across the bleeders and why 150mA for the new supply? A couple of mA over the bleeders should be sufficient, may cure your transformer buzz issue. If you are using the bleeders as a voltage adjustment, and you have a cap input supply, try a smaller input cap.

The new supply may be overkill in low ripple for the Aikido, but won't hurt anything.

Sheldon

[cbutterworth]

Sheldon,

The original reason for the bleeder resistors was to pull sufficient current to allow the chokes to function correctly (B+ / L = required minimum current). Of course, this is for choke input filters, but I reckoned that it certainly wouldn't hurt to use the same rationale for cap-input. Also, when I modelled a lower valued cap than 0.68uF, I got increased ripple.

It does seem that any cap-input filter does not really require a minimum load for subsequent chokes. My plan for tomorrow morning is to remove the bleeders and just draw current for the circuit. I expect that I'll end up with something like 340VDC loaded. It will certainly be interesting to see what happens to the buzzing then as the load will be lower.

As for the 150mA trafo, there are a couple of reasons for going higher (some of which may uncover my relative ignorance):

1. Doesn't higher current capability usually mean lower DCR of the secondaries? If so, then I believe that lower DCR is advantageous.

2. Given that my mains voltage varies from 118 to 124VAC, I think that a higher current trafo will be able to tolerate the higher voltages better and may be less prone to saturation - of course, I could be wrong on this. Also, this trafo will be physically larger and so is less likely to heat up significantly when my mains spits out 124VAC.

3. My four 6SN7 heaters require 0.6A @ @6.3V for each tube, I have four of them. The lower current Hammonds only have 2A 6.3V secondaries. The 150mA one is capable of 6.3V @ 5A which gives me plenty of headroom. At present, I have a separate 6.3VAC trafo on my Aikido. This could be removed and possibly replaced with a flywheel choke that Bruce Anderson really really likes.

Now, my chokes are rated at 150mA. Even though I will not be drawing 150mA, do I still need to ensure that my power trafo does not exceed 150mA in specs? I think that the answer to this question is "no" because it is the actual current draw that is critical and not the rated current draw.

ANother reason for going tube rectified, is that I think the rectifiers look amazing. Right now, I have an octal socket for the Amperex tube delay, which I can remove and use for the rectifier. Using something like the 5V4GA will result in an appreciable delay in achievement of B+ (10 seconds), which negates the need for the tube delay. The tube delay also uses an Omron relay for the actual B+ switching, and I have heard that these types of relays can themselves be a source of noise.

If anyone has any comments, please reply to this post. Would a 50mA trafo and separate 6.3VAC filament trafo be better than a single 150mA trafo that can provide B+ and 2.4A of filament current?

Charlie

[SY]

Would a 50mA trafo and separate 6.3VAC filament trafo be better than a single 150mA trafo that can provide B+ and 2.4A of filament current?

Yes.

Low DCR for a preamp transformer is useless. Worse than useless- it means higher ripple current. Any potential advantage of low DCR is nullified anyway if you use a tube rectifier.

Separating the heater transformer means that diode noise from the B+ supply can't get coupled back to the tube as a common mode signal on the heater line.

[cbutterworth]

Sy,

Great, so I can go with a lower priced lower current trafo. There are some Hammond trafo that put out 250VAC rather than 275VAC. After modelling a few in PSUDII, it seems that the lowest ripple occurs with a 5uF cap1 rather than 0.68uF. So, I should probably go with the lower voltage trafo and use 5uF to aim at a voltage of around 330VDC.

Some people use around 360VDC for their 6SN7's. Would the subsequent higher plate voltages improve dynamics?

Also how much noisier are tube rectifiers?

Charlie

[Sheldon]

The larger input cap with result in higher current spikes, which can generate more high frequency noise, so all things being equal, the smaller cap can make your overall job easier. This might be less an issue if you have a scope and can play with grounding and ps snubbing. Either way, you can get a good supply.

The tube rectifier shouldn't be noiser, unless it's a faulty tube.

Sheldon

[cbutterworth]

Sheldon,

So, I would probably opt for 275-0-275 @ 90mA for the trafo, and use a lower valued cap (0.68uF) to bring the voltage down to my preferred level? I could always add a bleeder resistor to increase current draw and bring B+ down further still.

Before I actually do this, Jim Hagermann suggested that I try taming my 124VAC mains voltage using a resistor in series with the transformer primary. He reckons that the trafo will be much happier with closer to 115VAC.

If this cheap fix works, than I may simply forego further modifications to the PSU, although I still do really, really like the look of tube rectifiers.

As it happens, I removed the 20K bleeders from the final cap of my current PSU. Removal of this 10-12 mA of draw means that the tubes are now seeing more like 320VDC. Interestingly, when I first built this PSU and measured it, I got a reading of around 360VDC. I get this reading initially, but after 10 seconds or so, the voltage begins to drop and remains at the 320VDC level. I suppose that being new to this kind of thing made me nervous about leaving the amp switched on while taking measurements. Furthermore, this also reassures me somewhat as PSUDII simulations come closer the 320VDC than 360VDC.


Charlie

[Sheldon]

I like the smaller input cap if you can make it work. The trade offs are lower voltage, higher ripple. Either way works, though.

No harm in trying a series resistor with the transformer primary. BTW, you can model this in PSUD by just adding the series resistance to the primary resistance.

Looks is as good a reason as any to use a tube rectifier. Other benefits are maybe less diode noise, gradual warm up. Downsides are more heat and higher transformer voltage required for the same output.

Unless you are preheating the amp filaments, the voltage drop you see after turn on would be expected based on the ramp up of current draw by the amp as its filaments heat up. You can model this too, by just plugging in a very low current draw for starting conditions. You should do this anyway, especially as you have removed the low value bleed resistors. I make sure all of the ps components, caps especially, are rated to take the output voltage of the ps with no load - which they may see in the event of a fault. Also analyse your circuit to make sure that components won't see voltages across them exceeding their ratings under the same fault conditions.

Sheldon

[cbutterworth]

Sheldon,

I am certainly going to give tube rectification a try. The last couple of weeks have seen me in a dilemma as to my next project. While I would love an amp for my office, the speakers would not be that good and my budget is low. I just spent a bunch on the Cornet2 phono pre and a turntable.

The other benefit from not doing a new tube project is that I can re-house my Aikido. After using quilted maple for my Cornet2, my wife would like something more than a plain pine chassis on the Aikido. We're thinking maybe two red/brown woods, darker for the front and rear, slightly lighter for the sides with nice dovetail joints at the corners (handmade this time!) and maybe a polished aluminum chassis rather than black.

Anyway, I'll try the series resistance on the primaries to reduce the mains voltage down to something more like 115VAC. Then, my only difficulty will be deciding when to switch over to tube recitifcation.

Thanks for your advice.

Charlie

[luvdunhill]

You can model this too, by just plugging in a very low current draw for starting conditions.

what values do you use here for the initial current and the delay? seems contradictory in some terms as wouldn't you have in-rush current at turn on? Finally, would a soft-start circuit using a resistance and a relay help alleviate these potential problems?

Thanks!