The Nemesis 813 PP Amplifier

[Nick]

Yep,

[izzy wizzy]

Converted the AC heated 813s to fixed bias and I think it might be a result. Well enough for me to live with. Yes, it hums a bit but it is small. I can hear it just from the listening position but stick a record on, even softly and I can't notice it. At normal levels, it isn't relevant.

Also, the hum that was there without HT, just filaments on, has gone completely. That I find very interesting.

So off now to convert the SMPS powered 813 filaments to fixed bias.

Cheers,
Stephen

[izzy wizzy]

Fixed bias would seem the way to go. Overall the SMPS one is quieter too but sounds a bit noisier due to the buzz than the all AC one. I'm tempted to convert it back to AC but will have a comparitive listen between channels first. I'll have to see long term if the AC one is quiet enough.

Good thing with fixed bias is way less heat. I imagine a seperate HT supply for the first stage would be a good thing so sometime over the holidays, I'll have to find some transformers in the shed.

At some point I want to try the MVs again but need more transformers for the 46s to free up the ones for the MVs. New Year for all that.

Cheers,
Stephen

[izzy wizzy]

I've put the SMPS powered fil channel back to AC and rearranged the humpot situation. Neither channel has audible 50Hz but I'm left with some 100Hz buzz. It's quite low and will do for now.

Sounding out the Joes got some suggestions and the one I liked most was what Silvercore do to power their 833a. They use two Meanwell RS-50-5 SMPS in series with the centre taken to ground. About £15ea from RS which is about the best price around. A Joe runs his on the bottem end of an efficient speaker and says they are silent. So if the buzz gets to me, that might be my route.

I also took Nick's advice about grid stoppers and now have none on G1 or G2. G1 has a ferrite; a multi loop one as that is what I had. I think this has helped too. I should do the same on the 46 to see what happens.

In chasing the buzz, I had the HT with 15H-330u-15u-70u. At one point I had it as 15H-70u and could not discern any more noise with my ear stuck in the speaker. Sound wise I should take a listen to see what I can hear. The last config had the L in the PSU connected via the RG213 cable to the 70u in the audio part.

Cheer,s
Stephen

[izzy wizzy]

Been puttng some thoughts together about the destination for this thing so I can plan where to go next and also to get some design input from my better half in order to make the thing safe and domestically acceptable. To give her something to design around, I have to have something visual.

Idea is to have it built in layers with a sleeve to cover it up with the design something along the lines of 50s/60s radio/console on casters.

I still need to be able to split the thing to work on it or transport it so the the top two layers split from the bottom.
The bottom layer is noisy HT power supply stuff. One HT for 813s, Tx-813-L and the other for the 46s, Tx-RG1250-L-C-L; 866 and RG1-250 MVs respectively at the front behind a screen/window. The external umbilicals are a pain and make the HT wiring very long so I plan to link the HTs to the upper layers vertically with 1000V rated 2mm test plug/sockets. The mains travelling vertically can be on something like regular mains connectors.

The middle layer contains the 813 filament transformers or SMPS directly under the 813s (shown at the back), 46 fil tx below the 46s with 46s HT cap below the IT(shown front left), 813 bias supply below the IT (shown front right)

The top layer is all audio built on a PCB point to point with the HT caps under the OPT with speaker conns next to it on top. Input sockets on top next to the input transformer. Pretty much as the current one is built but with better spacing. Plan is for it to be screwed under a routed recess in the top plate to stiffen the PCB. It is 18mm ply as is the bottom layer; middle layer 12mm ply. All held together with threaded rod in the corners.

Can't make up my mind about on/off switches. If they're on top at the back, handy, then I need two mains connectors to the bottom layer. If they're on the bottom layer as they are now, they are at the back which is kinda awkward but then again, they can't be accidently knocked on and then there is only one mains link to the top layer. Might be OK now but in future, they might not be accessible.

Any comments very welcome.

Cheers,
Stephen

[RhythMick]

Nice design, whats the software ?

Have you considered having switches offboard ? Reduces complexity, avoids nasty mainses near the audio and easier to reach for if something is wrong. With an obvious indicator to show power is applied.

[Mike H]

The external umbilicals are a pain and make the HT wiring very long

Shouldn't be a problem provided the supply capacitors are close to the amplifiers. I've used separate PSU's with leads a metre long or maybe more, without any issues. But in every case the [final] supply caps are in the amplifier section. HTH

[Nick]

The external umbilicals are a pain and make the HT wiring very long

Shouldn't be a problem provided the supply capacitors are close to the amplifiers. I've used separate PSU's with leads a metre long or maybe more, without any issues. But in every case the [final] supply caps are in the amplifier section. HTH

Assuming that the earth resistance from the metalwork and the earth pin on the IEC id 0.1 ohm and less and you can pass 12A from that IEC pin to the chassis without failure for 1 minute, then not a problem.

That does assume you want to comply with safety regs of course

Oh, and if the lead is removable then ensure that there is no dangerous voltage on any of the pins if the lead is removed when the unit is operating.

[Nick]

Sorry, I misremembered, that's 25A and two minutes in the UK.

[RhythMick]

Do you also need to put bleed resistors in both the amp and the PSU to ensure the caps can drain independently?

[Nick]

Do you also need to put bleed resistors in both the amp and the PSU to ensure the caps can drain independently?

You need to do both sides yes, but bleed resistors cant get the voltage down fast enough. Can't remember what the allowed time is but its in the order of a second I think I remember. Well, they could but you would be burning a lot of power when its working.

[izzy wizzy]

Nice design, whats the software ?

Have you considered having switches offboard ? Reduces complexity, avoids nasty mainses near the audio and easier to reach for if something is wrong. With an obvious indicator to show power is applied.

It's Sketchup 3D modelling software. Free. Use it for audio but mainly for house renovations work. Steepish learning curve, bit quirky but very powerful.

Not sure by what you mean by switched off board. I need two - one for filaments, the other HT. With switches at back either on top or back panel there's no mains anywhere near audio.

Cheers,
Stephen

[izzy wizzy]

Thanks for the comments.

I 'spose the umbilical thing is usually one of necessity. Things are too heavy so need splitting (this case) or too noisy to be close to audio bit say phono which is the case with mine (plus it's too big for one box) If I can do the connection internally, then I don't need an umbilical. Nick nighlights some of the issues but also you both have pointed out that it's OK too.

For the 813 HT, I'm thinking it being only LC with L in the bottom and the C in the top, can it be a good thing to have that 1000V pulsing DC going through a long lead? I dunno really. Currently it's RG213 coax through an N type RF connector. I also would like to try and keep the 1000V inside than travelling around outside; things can go wrong and it isn't something I want to go wrong. I hate planning for 1000V things going wrong.

The 46 HT is going to be LCL in the bottom and C in the top so not so bad. I have the RG213 and more N type conns so it isn't a cost thing. I was just thinking an internal connection with 2mm connectors would rid me of external connections and also reduce the length of the HT connections. I'll try it on the prototype to see if there's a difference - bit of fun

Cheers,
Stephen

[izzy wizzy]

For the 813 HT, I'm thinking it being only LC with L in the bottom and the C in the top, can it be a good thing to have that 1000V pulsing DC going through a long lead? I dunno really. Currently it's RG213 coax through an N type RF connector.

Actually it's not pulsing DC Not sure what I was thinking.
It's about 130mA of AC current so says PSUD. I should do more homework.

cheers,
Stephen

[izzy wizzy]

I've come back to this amp after a phono rebuild. I had also sounded out the Joes about my buzz/hum. Despite one person being able to get PP 813s silent on AC, the general consensus among anyone that has these things is DC is the way to go. Those that have 833 or 813 amps SETs use SMPS supplies. Someone pointed me at the Silvercore 833 and told me the power suppies it uses and he said his was silent.

So last week I bought 2 Meanwell 10A 5V supplies, connected in series with the centre tap taken to gound like Silvercore do. On first turn on, I was very disappointed as the valves didn't light up. Then slowly they came alive. It has a slow start. Woohoo.

And it is utterly silent. I can't quite believe it. After 15 years on and off, I finally have one silent amp. Running around like a looney, jumping up and down. Family think I've lost it. Now to buy another two.

I feel I can go forward now with other things and work towards a better build in a slightly smaller volume.

Cheers,
Stephen