Simplifying series filter capacitors

[cdemike]

I'm planning to revisit a Marshall-style build with a few mods, including a PT swap to get a big bump in B+. Right now the amp has caps rated for 500V, but I'm considering a PT rated at 200-0-200, which loaded should provide a B+ around 555VDC at idle. Since B+ at at the first two nodes would be increased higher than any commonly available electrolytic, I'm planning how to approach placing filter caps for the first two nodes in series. I'm specifically wondering if I might be able to simplify my setup versus the typical Marshall arrangement with 2x series capacitors at both the reservoir and screens nodes. Taking a page from Hiwatt power supply schematics, I'm wondering if there would be any disadvantage to this arrangement with 3 segments:

Capacitor question.PNG

My chassis is cut out for 2x cap style electrolytics, and I have board space for all the preamp filters, so I'd be implementing with one multi-section 50x50uF cap and one 220uF cap.

I'm concerned I'm missing something since I imagine Marshall would have cut costs by implementing something similar (1x fewer capacitors per amp seems like it'd be a worthwhile material cost savings) if the performance were equal. That said, this would require keeping a different value of capacitor on-hand, which I suppose could savings in a bulk order... Either way, not a concern that applies to my hobby build.

[Stevem]

Keep in mind that a voltage doubling PS like in a Highwatt also only provides 1/2 the current output that the B+ winding of the PT can provide.

[Helmholtz]

Keep in mind that a voltage doubling PS like in a Highwatt also only provides 1/2 the current output that the B+ winding of the PT can provide.

Why 1/2 the current?
I think he means a bridge rectifier connected across the full HT winding (i.e. 400V) with the CT used to balance cap voltages.
This is just a full wave bridge rectifier circuit, where the winding current is 1.6 times the DC current.
As there's no net current in the CT, it doesn't change anything. The rectifier circuit works the same no matter if the CT is connected or not.
Calling this circuit a voltage doubler is misleading.

[cdemike]

Fair point. Currently the amp has a Mojotone JTM45 PT which is set up with full-wave rectification and grounded CT, and it idles at 434VDC with 50mA through 2x KT88. So I'm estimating power consumption starting with 2 x 0.05 x 434 = 43.4 watts. Accounting for preamp consumption by looking at voltage drop across the dropping resistor between the screens and PI nodes, that's closer to 45 watts. I like the amount of power supply sag I have now, so to keep an equivalent amount of sag when switching from full-wave to full-wave bridge rectification, I should a PT rated for approx 90W or perhaps slightly higher given the higher B+, right?

I was considering this PT before the rectification wrinkle: https://www.antekinc.com/content/AS-05TC200.pdf

Factoring in the differences in rectification, this strikes me as the better option while providing B+ in a similar range at the original target: https://www.antekinc.com/content/AS-1T230b.pdf

The datasheet describes 208V at 500mA. Since the secondaries would be wired in series rather than parallel as in the datasheet, that would translate to 250mA. Factoring 62% available current via FWBR vs full-wave (going off the Hammond rectifier guide), that would be 208V at 155mA reaching the rectifier, i.e., 31.47W. Since the estimated current demand would be relatively similar, the estimated current would still be approx 45w, so 1.43x higher. Current drops from 233VAC to 208VAC under those loading conditions in the Antek datasheet, so the voltage drop under load would be 1.43 x 25 = 35.75V, producing an estimated HT secondary voltage of 197.25VAC across each secondary or 394.5VAC reaching the rectifier. So estimated idle B+ would be 1.414 x 394.5 = 557.8VDC. Does this math seem right?

Also I realized it may be better to show a picture of the whole B+ string. This is what I was considering:

Capacitor question part 2.png

Versus this being the "stock" setup:

Capacitor question part 2a.png

Edit: caught a mistake in the first attachment.

[cdemike]

Keep in mind that a voltage doubling PS like in a Highwatt also only provides 1/2 the current output that the B+ winding of the PT can provide.

Why 1/2 the current?
I think he means a bridge rectifier connected across the full HT winding (i.e. 400V) with the CT used to balance cap voltages.
This is just a full wave bridge rectifier circuit, where the winding current is 1.6 times the DC current.
As there's no net current in the CT, it doesn't change anything. The rectifier circuit works the same no matter if the CT is connected or not.
Calling this circuit a voltage doubler is misleading.

Welp, now I'm confused. I was under the impression that a FWBR has 0.62x less current capacity than a full-wave rectification setup? At least that's how I have been interpreting this: https://www.hammfg.com/electronics/tran ... /rectifier

[Helmholtz]

I was under the impression that a FWBR has 0.62x less current capacity than a full-wave rectification setup?

If your amp has a DC current demand of Idc at full output, the winding RMS current will be
- 1.1 Idc with a full wave (2 diode) rectifier,
- 1.6 Idc with a bridge rectifier.
Actual difference is sqrt(2) or 1.414.
Typical Idc values are 250mA for a 50W amp and 500mA for a 100W amp.

The Hammond sheet might have flaws.

[cdemike]

I was under the impression that a FWBR has 0.62x less current capacity than a full-wave rectification setup?

If your amp has a DC current demand of Idc at full output, the winding RMS current will be
- 1.1 Idc with a full wave (2 diode) rectifier,
- 1.6 Idc with a bridge rectifier.
Actual difference is sqrt(2) or 1.414.
Typical Idc values are 250mA for a 50W amp and 500mA for a 100W amp.

The Hammond sheet might have flaws.

Thanks so much. If I'm following, it seems like the calculations regarding PT should be right -- the 100W transformer with a FWBR and correspondingly higher B+ should provide comparable sag to the 45W PT that's there now, right?

And would substituting the 220uF for the 2x 100uF in the "bottom" parts of those series capacitor arrangements make sense?

[Helmholtz]

Generally a PT should have a VA rating of at least twice the output power.