Kay 703-based design

[oxbow_lake]

Hello!

Wanting to build an amp that is simple enough for me to try point-to-point wiring but is something other than a Champ clone, I modified the circuit for a Kay 703 and built it. The schematic of the original amp is below.

PXL_20231017_095534752.jpg

I modified it to use solid state rectification and replaced the tone circuit with another (still simple) one based on a Gibsonette. The series filaments are still run off the mains, though with a big resistor to replace the rectifier filament. The rest of the circuit is behind the isolation transformer, and the amp has a proper safety ground. This modified circuit is below.

703 filament resistor.jpg

As built, the tonal quality is surprisingly good. Good enough to warrant the effort of working out the kinks. The issue is that there is an unacceptable amount of 120 Hz hum. It's really quite loud. This hum is present when the input jack is grounded, and also when the volume is all the way off (grounding the 50L6 grid).

At this point I will swallow my pride and attach photos of my build. Be kind, I'm trying to learn. And ignore that hideous fuse holder.

PXL_20231017_103346167.jpg

PXL_20231017_103436084.jpg

From left to right on the panel it's input jack, tone, volume, external speaker jack, power switch.

Moving the output transformer around does not alter the hum volume at all. Moving the OPT wires around within the chassis (blue/red twisted pair) does not alter the hum volume at all. In terms of layout, I tried to keep to more-or-less a star ground configuration, coming to a chassis-mounted lug near the input jack. The jacks are not isolated so it's far from perfect. There is also a chassis ground near the power cord where the safety ground connects. Beyond that I just tried to keep the leads short, and the power supply stuff relatively far away from the signal stuff. Sorry, I know asking anyone to look at homebrew point-to-point wiring is probably not very polite.

You folks were super helpful with debugging my last build, so I thought I'd ask. Do you think the hum is something inherent to this simplistic design (i.e. lack of power supply filtering), is it more likely a problem with my layout, or something else? Sadly I don't have an original 703 to compare to!

Elsewhere on the internet I saw someone build a similar amp using a step-down transformer to provide 60V, then a voltage multiplier to give B+. This is pretty cool, and allows the filaments to be isolated from the mains as well as getting rid of that huge filament resistor (which gets scorching hot). Furthermore, it would be easy to add more power supply filtering which may help. I mocked up a schematic, and it's easy enough to build, but it'd be nice to know if there's an actual design problem with the circuit as I currently have it.

703 voltage doubler.jpg

Thanks everyone.

[sluckey]

Adding another filter stage before your plate node will help cut down the hum considerably. If you can replace R8 in your revised schematic with a choke it will help even more.

I have a very similar amp built with the suggestions I just made and the hum is quite low.

http://sluckeyamps.com/smoky/smoky.pdf

Will your PT handle the extra 150mA current needed by your filaments in the revised schematic?

[statorvane]

Are you using a 1/2 wave rectifier in the amp? Might be a source of hum.

[oxbow_lake]

That Triad F-59X is rated for 24 VA, 400mA. The filaments will take 9W. The plate/screen currents of both tubes add up to maybe 65mA total (probably not actually close to that), so running at 160V let's say 10W? So if I did that math right then the PT should be able to handle it.

The rectifier in my current design is a full-wave bridge rectifier.

The voltage doubler as drawn is half-wave, but I could implement it as full-wave, like this? I'm not an engineer, feel free to correct/make suggestions. Maybe balancing resistors over the doubling capacitors.

703 full wave doubler.jpg

One question about the voltage doubler - the first one I drew has that 68R resistor between the filaments and the diodes. Merlin's doubling circuits don't have this. I copied it from a schematic I found online. Is it necessary? Is it just helping to isolate the filaments from the rest of the circuit? I can't find the schematic I ripped off anymore, or I'd link it for reference.

Thanks again!

[johnnyreece]

Were you thinking of this?

http://www.kirtland.com/guitar1/guitar1 ... _Amp-1.gif

[oxbow_lake]

Yeah, that's it. Thanks!

[oxbow_lake]

Here's an update nobody asked for.

I built it with the voltage doubler. The hum is significantly less, though it's far from hum-free. It runs *much* cooler without that huge filament resistor.

In the spirit of this being a budget amp, I crammed the thing into a little Squier SP-10 practice amp cabinet. It looks like hell, but it was fun to build. In case anyone wants to know what it sounds like I attached an audio clip, pardon my playing. In the clip I'm using a greenback external speaker instead of the tiny crappy Squier 6" internal speaker. Just a reverb pedal for effects.

The final schematic is here, along with a layout though I wouldn't really recommend anybody copy that!

Cheers.

[Phil_S]

Nobody asked me either, LOL. FWIW I like it! Well done!

[TUBEDUDE]

Here's an update nobody asked for.

I built it with the voltage doubler. The hum is significantly less, though it's far from hum-free. It runs *much* cooler without that huge filament resistor.

In the spirit of this being a budget amp, I crammed the thing into a little Squier SP-10 practice amp cabinet. It looks like hell, but it was fun to build. In case anyone wants to know what it sounds like I attached an audio clip, pardon my playing. In the clip I'm using a greenback external speaker instead of the tiny crappy Squier 6" internal speaker. Just a reverb pedal for effects.

The final schematic is here, along with a layout though I wouldn't really recommend anybody copy that!

Cheers.

Good job, sounds nice. It's good to save it from the landfill.

[didit]

Hello --

Oh, Kay. Admirable. The only remaining hum suppression tweak would be to swap in a choke for the large power resistor (R7) before B+1. I took a quick 10 minute detour and roughly simulated with a 2H/57R choke, which suggests reduction down to 25-30% of B+1 ripple. That may however elevate above your "budget" category.

..

[oxbow_lake]

I've never used a choke, so I'm willing to spend ~$20 to try something new.

At Amplified Parts I see a couple of options in that price range - either 1.5H/56R or 3H/86R. Same price for either. Would the slightly lower cutoff frequency warrant the slightly higher voltage drop over the 3H choke? Do I need to worry about LC resonance, or is that probably not a big deal?

What software did you use to simulate the circuit?

Lastly, do I need to worry about physical layout for a choke, or can I more or less squeeze it onto the chassis wherever it will fit?

Thanks!

[didit]

Hello --

I've never used a choke, so I'm willing to spend ~$20 to try something new.

At Amplified Parts I see a couple of options in that price range - either 1.5H/56R or 3H/86R. Same price for either. Would the slightly lower cutoff frequency warrant the slightly higher voltage drop over the 3H choke? Do I need to worry about LC resonance, or is that probably not a big deal?

What software did you use to simulate the circuit?

Lastly, do I need to worry about physical layout for a choke, or can I more or less squeeze it onto the chassis wherever it will fit?

Thanks!

A choke will certainly have a bit of dynamic magnetic flux but it's low by comparison. Should be OK assuming you're sensible and avoid proximity to stuff most sensitive to noise.

Simulation was quick rough approximation using https://www.duncanamps.com/psud2/. Results though are consistent with experience comparing CLC pi filter vs CRC.

One can certainly find a suitable choke from this set https://www.amplifiedparts.com/products ... en-bracket. Voltage is low, so not a constraint. Measure current capacity required as it's a factor. Anticipate less than 100mA but by how much? Recommending a quick engineering detour estimating your resonance frequency . If memory serves (though often doesn't) parameters from 156M were used in my quick kludgy simulation run. Henrys vs resistive voltage drop is certainly a trade off.

Best ..

[oxbow_lake]

I went with the 3H choke. Pretty quiet! Thanks for the help.

[didit]

I went with the 3H choke. Pretty quiet! Thanks for the help.

Good to hear it went well.