The Newton 300B

[Thermionic Idler]

I've just looked up the Maida regulator (I presume we're talking about this one?). The use case is perfect, I need to drop about 25 volts at 20.5mA, so I expect heatsinking to the chassis will be enough.

Yes, that's the one, however, the design was originally published on DIY Audio, so public domain, and I purchased a batch of boards produced by a DIY Audio member that incorprate the rectifier diodes and filter cap - I have several spare boards, populated with smd parts, is one of interest? They're what I use in my Masterpiece HPA and SE-OTL amp input/driver stages. I also have a calculator spreadsheet, which makes life easier with working out the parts required to set the voltage/current parameters for a specific use case.

They may well be - what are the dimensions of the boards as I'm limited to about 57-58 mm width between the input transformer and one of the output stage caps. The Neurochrome ones will *just* fit...

[Ray P]

The ones I use are larger than the Maida ones Dave

https://www.diyaudio.com/forums/swap-me ... hingy.html

[Ray P]

https://www.diyaudio.com/forums/vendor- ... lator.html

[Thermionic Idler]

Yeah - I suspect it won't fit in the available space.

The redundant 80uF cap attaches to the back panel by means of a large stud bolt. My cunning plan to avoid having to drill new holes in the chassis, is to have FPE make up a couple of raw aluminium adaptor parts to attach the PCB via the same hole, and heatsink Q1 to the back panel - but there's literally only 2 or 3 mm to spare!

[Ray P]

I'm not even sure you'll need the heatsink Dave as you're not asking the regulator to deliver much current I guess?

If you tell me your required output voltage and current I'll plug the numbers into the spreadsheet if you want - will give you heatsinking requirments.

Somwhere here I have a new set of the plastic film caps (the big blue ones) for the Neurochrome maida board - I'll find them later and stick them in with the 12SN7s.

[pre65]

Yeah - I suspect it won't fit in the available space.

is to have FPE make up a couple of raw aluminium adaptor parts to attach the PCB via the same hole, and heatsink Q1 to the back panel -

Does it need to be anything fancy ?

If not, can you do a sketch of what it needs to look like ?

[Ray P]

I found them so have C1, C2 and C4 (and R4) covered for you if you go the Neurochrome Maida route Dave.

[Thermionic Idler]

I'm not even sure you'll need the heatsink Dave as you're not asking the regulator to deliver much current I guess?

If you tell me your required output voltage and current I'll plug the numbers into the spreadsheet if you want - will give you heatsinking requirments.

Somwhere here I have a new set of the plastic film caps (the big blue ones) for the Neurochrome maida board - I'll find them later and stick them in with the 12SN7s.

Hi Ray - well according to the Spice model (reality may vary) the input voltage will be 432V, required output 407V, so a 25V drop, with the current being circa 20.5mA. I make that about half a watt dissipated through the regular and it should give me a fairly textbook operating point for the 12SN7's.

I found them so have C1, C2 and C4 (and R4) covered for you if you go the Neurochrome Maida route Dave.

Splendid, thanks, will very likely take you up on that offer.

Yeah - I suspect it won't fit in the available space.

is to have FPE make up a couple of raw aluminium adaptor parts to attach the PCB via the same hole, and heatsink Q1 to the back panel -

Does it need to be anything fancy ?

If not, can you do a sketch of what it needs to look like ?

It might need to be fancy, I haven't decided the orientation yet though. I'll be doing the plan in CAD anyway so I'll share the diagrams when I've got them drawn up.

EDIT - actually if I don't need the heatsink, or only need the small type that attaches to the device itself, that makes the mounting requirements a lot easier. Just an L-shaped piece of aluminium with a large 13mm hole on the right angled bit, and 4 holes for the PCB on the other.

[Ray P]

Hi Ray - well according to the Spice model (reality may vary) the input voltage will be 432V, required output 407V, so a 25V drop, with the current being circa 20.5mA. I make that about half a watt dissipated through the regular and it should give me a fairly textbook operating point for the 12SN7's.

Here's what the neurochrome calculator says;

Capture.JPG (196.36 KiB) Viewed 4413 times

[pre65]

It might need to be fancy, I haven't decided the orientation yet though. I'll be doing the plan in CAD anyway so I'll share the diagrams when I've got them drawn up.

EDIT - actually if I don't need the heatsink, or only need the small type that attaches to the device itself, that makes the mounting requirements a lot easier. Just an L-shaped piece of aluminium with a large 13mm hole on the right angled bit, and 4 holes for the PCB on the other.

Awaiting diagram.

[Thermionic Idler]

Thanks guys.

[Thermionic Idler]

Some more tweaks made today to add to the build log.

Firstly the feet - they are fixed on with plastic bolts to ensure that they are insulated from the chassis interior, so I couldn't do up the M4 nuts too tightly otherwise the threads would strip. Unfortunately, the effects of a house move followed by a trip to Hull when I took them to a meet at Steve's, had caused the bolts to loosen. One of the feet came adrift upon arrival in Hull. They now have flanged locknuts holding them in place to stop that from happening.

Next - GP10Y diodes in series with the rectifiers (actually just before the fuses) to help the rectifiers last a bit longer and reduce the potential for arcing and other nasties.

I then made some further changes to tweak the operating point on the 300B's to somewhere in the region of 350V / 60mA, or at least a voltage / current ratio in that ballpark - why second-guess Western Electric. I started out really cooking these 300B's (per Herb Reichert's recommendation), but found that each time I cooled off the power dissipation a bit, it sounded better. After lots of calculations a couple of months back, I removed the series resistance on the output stage rectifier (the 5U4 doesn't need it), and changed the cathode resistance on the 300B pair to 620R from about 452R. A 12W Mills 120R resistor was placed in series with the existing 500R 50W resistor attached to the heatsink, with the previous paralleled 4.7k resistor being removed.

On channel A, following these changes the plate voltage measured 360V with 61.39mA of current, and on channel B 369V with 59.77mA. I'm calling that close enough. The plate dissipation has now gone down to 22W per valve, so that means I can run mesh plate 300B's if I want to (not that I can currently afford them!) and I should get a good long service life, even these Chinese 300B's aren't cheap!

Having had a good listen after the changes were made, I can't hear the diode modification at all so that's good. And I think the operating point changes have kicked up the transparency a notch, certainly nothing has degraded at all. I suspect peak power will be down slightly but we don't listen at ear-splitting volumes anyway.

The next thing to look at will be the input stage - I intentionally left that untouched today so that I could hear and assess the effect of altering the output stage bias point. I want to increase the current on the 6V6's a bit, but reduce the 12SN7 power dissipation to 1.8W total or thereabouts, as again I don't think that valve needs to run as hot as it does currently. That'll allow me to run any 12SN7, not just the GTA-designated ones.

I'll be glad when these tweaks are all done, because high voltage stuff bloody terrifies me.

[Ant]

I'll be glad when these tweaks are all done, because high voltage stuff bloody terrifies me.

Thats why i dont bother any more
Im dreading rebuilding eds sv811 into a case even though it does sound fab.
I actually only need 2 bits of nice timber for the side cheeks for the case for it, i just cant seem to make a start

[Thermionic Idler]

I'll be glad when these tweaks are all done, because high voltage stuff bloody terrifies me.

Thats why i dont bother any more
Im dreading rebuilding eds sv811 into a case even though it does sound fab.
I actually only need 2 bits of nice timber for the side cheeks for the case for it, i just cant seem to make a start

I hear ya. I've been putting this job off for months, first I have to haul the things up and down the stairs on a sack trolley because they are so heavy, trying not to knock into anything and spoil your lovely casework. The actual soldering work I do enjoy, but then comes time to get a fresh set of voltage readouts and my blood always runs a bit cold at that point.

I've had these thick electrician's gloves for years, to be honest I find them too unwieldy when it comes to holding the meter probe so I just exercise great care, but recently I've taken to wearing one of them on my left hand as a buffer against stupidly "completing the fatal circuit" as it were. Trying to hold one hand behind my back or in my pocket just didn't work. And I figure this is all a good thing, dealing with this stuff needs to frighten me otherwise I wouldn't be careful enough.

At least the power supply has relatively little capacitance in it, and there are a string of 12k resistors across it which means the PS caps discharge in seconds after switch-off, so once the mains cable is yanked I can breathe easily.

But yeah, one thing I'm loving about the Silentswitcher project is the prospect of being able to test things out on the bench without that 450V Sword of Damocles over me.

[simon]

I found crocodile clips quite useful and have saved my sanity a few times. Connect a clip to the point of interest with the amp off, and the other end to your meter. Power up from a safe distance.

It's not ideal for having a quick poke around a few places, but for just a few places I've found it easier to read the remote end of the Croc clip than poking around inside the guts. Even just having the earth probe connected really helps. You might already do this in which case please ignore!