Alternative take on the Radiotron SE2A3

Almost there. Just the grid caps to wire before I plug in and test.

Took some trial and error to sort the input - grid caps and thin screened wires are a pain. This is where I am up to:

Almost_there_interior.jpg

Decided to connect 0V signal to chassis near the RCA. Also grounded the grid wire screen at the same point. The two binding post thingies you can see (green wire attached) are not speaker outputs - they are banana test points for measuring 6C6 cathode current (voltage dropped across a 10 Ohm R) externally so I can set 6C6 plate and output grid voltages.

Connector mounting plates all installed:


Almost_there_exterior.jpg

The mounting plates are inset to provide protection for switches, fuse holders, and connectors. You can see the banana test points next to the RCA - multimeter pugs in there.

Should be able to sort the grid caps tomorrow and maybe even fire up for a smoke test and see where (if!) the voltages settle. I'm a little apprehensive - tried quite a few new techniques in this build and really not sure if it will work.

Oh, and they weight a heap - about 20kg per channel is my guess... 10kg per Watt. Daft. Looking forward to a simpler next build - maybe EML 20B/M drives PP 45?

Cheers.
 
I have to say the way you've laid things out is very impressive. The graceful arcs of the wires and the way you've secure the components is very cool. No lack of room anywhere to get in the way of soldering. Probably cuts down on interference or noise entering. Nice job.
 
I have to say the way you've laid things out is very impressive. The graceful arcs of the wires and the way you've secure the components is very cool. No lack of room anywhere to get in the way of soldering. Probably cuts down on interference or noise entering. Nice job.

Thanks muchly for your kind words. I had a few frustrating moments of difficult access to some solder points, but overall quite good. The problem will come if I need to change out some of those resistors mounted through nylon posts that are supporting capacitors - it can be done, but it won't be easy.

Regarding component mounting: I did that so I could experiment with lead-to-lead connections and distance components from each other and the chassis while still maintaining direct routing, accessibility and reasonable size.

Next build will be more compact, elegant and lighter.

Should, finally, get to the grid caps this evening - the delay has allowed me to think of a more reliable connection to the Yamamoto caps... here's hoping!
 
Well, grid caps are done with a really solid, reliable connection, though there is 4cm of wire that is not screened - should be fine. Didn't cover the connections to make 'em look pretty but they are functionally fine.

No photos - it's getting late and I don't have a decent internet connection (ISP outage).

Will plug them in and start testing tomorrow - hopefully they work and voltages settle roughly per the design.

Cheers.
 
Last edited:
Up and running! Still tethering to my mobile for internet - no photos until weekend.

Voltages mostly fine. 2A3 filaments are higher voltage than I'd like (but still within tolerance) so I'll grab some dropper resistors in the next week. Screen is also a bit high at 82V, but well within acceptable operating range.

Hum is much lower than previous amp - JJs are in the amp now and they are my noisiest tubes: left amp is < 1.5mv, right is < 2.5mv. Ear has to be very close to speaker to hear noise. There is a slight bee-like buzzing in both channels which I can hear when directly on axis of the horn and my ear is close - oscillation? Your thoughts? Also, the expensive, over-specified power transformer whines loud enough to hear at the listening position on quiet music; I can't feel any vibration and did not notice the sound when the amps were on the bench for testing - any ideas?

I'll let it settle in and clean up the AC cables and general wiring, change tubes etc and see if the noise improves. My guess is it will not, so any suggestions much appreciated.

Cheers.
 
Last edited:
Bee like buzzing is usually noise coming from the high voltage supply. A practical test to diagnose this is to get your clip leads out and add 100uF to the second node in your power supply, then listen/measure, then repeat as necessary.

If you have an oscilloscope, you can look at the noise on your HV rail to verify that this is where the issue is. If you have a really fancy scope, you can put a very steep high pass filter in at about 75Hz to knock out the heater hum, then you'll get a better look at power supply noise.

Who wound the power transformer? The usually noise from a power transformer is a mechanical vibration that sounds a bit like hum. To get rid of this, tightening mounting hardware is always my go to option, but since your transformer is in a shroud, that may not be an option.
 
Bee like buzzing is usually noise coming from the high voltage supply. A practical test to diagnose this is to get your clip leads out and add 100uF to the second node in your power supply, then listen/measure, then repeat as necessary.

If you have an oscilloscope, you can look at the noise on your HV rail to verify that this is where the issue is. If you have a really fancy scope, you can put a very steep high pass filter in at about 75Hz to knock out the heater hum, then you'll get a better look at power supply noise.

Who wound the power transformer? The usually noise from a power transformer is a mechanical vibration that sounds a bit like hum. To get rid of this, tightening mounting hardware is always my go to option, but since your transformer is in a shroud, that may not be an option.

Thanks Paul.

I tried some new techniques with this build - Kelvin connected PS capacitors, screen bypass grounded to the top of the cathode, direct coupling etc... also decided not to add the 220 Ohm 2A3 grid-stopper. Without help, hard to know where to start looking at what might be causing the issues (I had none of these with my previous build).

I'll try adding the 100uF to the PS node - I assume I'll need Kelvin connected caps like are already in the supply. The PS modelled with low ripple, but hey, no substitute for reality! No scope here, but I've been meaning to get something set up so I'll look into that in the coming weeks.

The power transformer is one of the more costly custom jobs from J&K Audio in Malaysia - generally well-regarded. It looks well made and should be coasting how I'm running it... I'll tighten the mounts and maybe add some damping. It may have not been making noise on the bench because the transformer would have been resting on its side on a book - probably damping any vibration...

I have faith that you'll figure it out and the amps will sing wonderfully.

Thanks for your confidence in me! It is sounding fine now, but I'd like to sort the mechanical whine.

Cheers.
 
Last edited:
I "hear" you. My amps are sounding astounding to me but it is hard to ignore the hum once you know it is there.
 
What is "Kelvin" connected cap?

Kelvin capacitors have four connection: -ve in, -ve out, +ve in, and +ve out. Apparently this results less high-frequency noise than standard 2-lead connections. I like that it forces you to connect to the cap in a way that it becomes more difficult to blend PS and signal currents.
 
Had some time listen in a less critical, more relaxed way at low to moderate volumes. It (I?) will still be breaking in, but it sounds very, very nice. Similar balance, tonal density and musical flow to its progenitor, but different also. Music is simply present and expressive; does not seem to be "reproduced".

I changed rectifiers as a long shot to see if that improved the mechanical whine from the transformers (it didn't), but they do make a slight difference to the sound: the '57 Sylvania 5931 are linear, resolved and snappy; RCA black plate 5U4G are weightier and gentler. The amp remains so very listenable with either - I can't imagine losing much sleep over which I use.

Regarding the whine, I'll damp and tighten the transformers shortly and contact the winder. I measured 0.8VDC riding on the mains (using a calibrated Agilent RMS meter) - not sure if that could contribute to the issue?

Cheers.
 
Last edited:
DC on the mains could cause problems. There are a few solutions to getting that off the line that are inexpensive experiments.
 
Thanks Paul - I'll look into it. The damping may have helped a bit also - the noise is present, but not as intrusive as before.

Right channel - needs a clean:
needs_a_clean.jpg

Cheers.
 
Well, this amp's ability to immolate expensive output tubes forces a redesign. Paul provided some good advice on how to retain direct-coupling, but I reckon I'll just implement classic RC coupling. I included the necessary secondary on the main transformer for that purpose, just in case. May even switch to DC filaments on the output tubes - line voltage fluctuation and differing noise performance of DHTs is a bit of a pain.

So, I'll try to keep as much of the amp as possible, but a bit of design work required - hopefully the end result will be decent.
 
I wouldn't blame yourself or the amp for this. There are plenty of 2A3s that are going to be fine with that circuit design, you just don't happen to be using any of them.
 
Thanks again Paul - I really appreciate your contributions, assistance and encouraging words.

It was a learning experience for sure... and I'll still end up with a nice amp. My decision now is how far I go with the redesign and build: 1) make the minimum changes to make it work; 2) moderate changes to address high impact/low difficulty issues; or 3) essentially full rebuild to fix some of the things I'm not 100% happy with. Option 2 is looking appealing right now - with my next amp being the one on which I can exert my perfectionism. :)

Thanks again.
 
Last edited:
I have a couple of nice 2A3 pairs and don't want to be limited in future. I've ordered the parts to convert to rc-couple.

I have been playing the amp and have been through about 10 switch on/off cycles. I had no issues with a Shuggie 2A3C (first time I've heard that tube sound decent), but have had an issue with the JJs: a nasty crackle at about the time B+ would be ramping up on the tubes. So, yes Paul, some tubes are not suitable.

It really does sound great... soon I'll know how much greater it sounds than the RC-coupled equivalent (using almost the same components). More learning.

I'd like to learn more about the issue and why some folks claim no issues running EMLs and JJs in similar topologies (Garber, Serious Stereo etc.). Any insights or further explanation would be most helpful. That said, I probably won't build this topology again - monkey (driver current derived from output cathode) or transformer coupling in future... maybe even an all DHT PP.

Cheers.
 
If you look at the Loftin-White 2A3 design (Garber), there is a string of resistors from B+ to ground, and the driver stage plate load resistor comes off of one of these nodes. That serves to limit the B+ seen at the grid on turn-on. Also, the LW design pre-dates all the "new" 2A3s that don't work well in DC amps. The SS amps use slow turn-on tube rectifiers (I think) which will mitigate the issue a bit as well.

You can also try putting a UF4007 diode between 2A3 grid and hum pot center tap on each 2A3 with the banded end toward the hum pot. When you turn the amp on, there will be a path to ground through the driver plate load resistor, diode, and cathode resistor. This will keep the grid voltage on the 2A3 down a bit.
 
Back
Top