• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

#26 pre amp

Andy,

I always prefer(red) low static resistance in choke/IT/OPT designs. This usually results in using thicker magnet-wire and bigger cores, which I like wrt lo bass-response..
Core material is also of big influence on sound (and performance)...
Interesting that you prefer the smaller Hammonds.. ;)

Oh, and I often use a low value resistor between plate and choke to "tune" the sound...

Rob
 
If you are in Europe the cheapest place is probably Bluebell Audio. Philip, the owner, is very into tube audio and is a big fan of the Hammond broadcast transformers. He does good deals if you buy a bunch of stuff, and anyway his prices on most stuff are cheaper than anybody else in Europe by some margin. Recommended! He's been selling the 156Cs in bunches of four. Apparently Hammond has increased production of them 200% in the last year and were surprised that so many were going out the door.

Bluebell Audio

Andy
 
I just follow Gary Pimm - he measured the responses with a scope. So bottom to bottom with the wires coming out the same side, pin 1 to pin 1. I assume I've understood this correctly. If I haven't, let me know!

As for whether to ground the chokes - good question!! Can we have some views? The chokes are dipped in some varnish stuff and you have to strip down to the metal to make contact.

I have reduced the choke input power supply in my 27 line stage to a 150Hy Hammond 156C and a Solen 10uf capacitor per channel. Still clean and quiet. I will be adding a table of values to the web site soon. One trick I did to get more bandwidth from them was use 2 in series. This doubles the inductance and halfs the capasatance. Frequency response went from 16-30K with one 156C to 9-80K with 2 in series. The high frequency went up more than double because there was a dip at 50Khz that was -4db with 1 choke and it was only -2db with 2 chokes. With 2 in series the frequency response was -2db at 50K rising to -1db at 60Khz then rolling off to -3db at around 80Khz. The self resonance is very low Q, a gradual hump between 50K and 70K that is only 1-1.5db high. Proabably low Q because the DC resistance is almost 3K. I varied the current from 5ma to 10ma and didn't see any frequency response varaition. Could have been the plate resistance decreased with the increase in current and counteracted the loss of inductance caused by the higher current. Another trick that you can do with 2 chokes is mount them bottom to bottom on standoffs so one choke is "right side up" and the other is "upside down" and wire them out of phase. This makes the whole setup "hum bucking". I measured the hum rejection at 20db as compared to both chokes mounted in the normal way. The test was having my Weller soldering station sitting 1 foot from the line stage. Turning one choke over dropped the induced 60Hz voltage by a factor of 10. Gary Pimm


Nice sounding, especially for the money! Use 2 in series. This has 3 things going for it.
- 1st, the inductance is double so you get 1 more octave on the bottom end of the frequency range.
- 2nd, because the capacitance of the chokes acts as 2 caps in series you will get better high frequency response. In the 26 preamp I built the 3dB down point went from somewhere near 30K to out past 70K. The single choke had a 4dB dip around 40K. With the pair of chokes the dip was only 2dB. There is a broad very low Q resonance around 60K so with 2 chokes in series it drops 2dB down near 30K, then rises back up gently and finally drops below the -3dB point somewhere above 70K. The numbers are vague cause I'm working from memory of a project done several years ago...
- 3rd, you can mount the chokes on standoffs, bottom to bottom and wire the chokes with one winding backwards. This makes the pair of chokes hum buckers. In my test setup the pair of chokes picked up 20dB less hum than a single choke. Gary
 
Hum/buzz on 26 preamp

I hope someone can help with this. I have a 26 pre with a loud hum/buzz. The 26’s are resistor loaded and LED biased, with Rod’s boards supplying the heaters. 50K stepped attenuator, star earthing in amp box, PSU connected via 2 umbilicals (1 for B+ and 1 for DC to Rod’s boards).
The PSU has a 100VA EI transformer with screen and HV secondary + 2 x 6.3V 3A secondaries; hybrid bridge with a dual damper diode (6BY5GA). The screen is connected to 0V. Smoothing is fairly heroic - for B+ it is clcrcrc (7H choke, 1st C is polyprop 5uF and others are 330uF) and to Rod’s boards it is crcrcrc (large value caps). By all accounts and according to PSUD, ripple should be almost gone in both supplies.

I got a slight improvement by snubbing the SS diodes in the bridge, so suspected that the problem had to do with the bridge. The heater draw of the 6by5 is 1.6A, which is quite a lot of ac current! Suspected it might be coupling somehow, so I tried supplying the rectifier valve’s heater with DC from a switch mode power supply. The frequencies within the noise were slightly different, but the noise was even louder.
I suspect that the problem has something to do with the PSU (which is quite compact). I hope it is not transformer problem.
Anyone got any bright ideas? All help much appreciated.
Best regards,
William.
 
Last edited:
Hi William, nothing about your construction stands out as being hum-inducing.

Maybe we are down to some feature of the layout. First, may we assume that the rectifier heater wires are twisted, and don't pass to close to other parts? Rectifier heater connected to HT+ or GND?

Is the first 330uF in the PSU chassis? (should be).

How does the PSU chassis connect to the amp chassis, and how does it connect to mains safety earth? How/where does the PSU B+ negative connect to chassis?

Layout photos can often reveal odd problems.
 
Rectifier heater wires are twisted. I tried it floating. It is connected to ground currently (with external DC sujpply).

I did test the diodes. I can't remember but I think I replaced them at one point. I did find snubbing them helped a bit, but not much. Maybe I should replace them again, just to be sure?

All smoothing is in PSU chassis, with extra decoupling caps where B+ enters amp and again right at plate load R's. Decoupling caps at input are right at input socket, between B+ and 0v. Decoupling caps at Plate R's connect to star earth (copper plated board) which connects to 0v by wire to socket. Heater supply is not connected to directly to 0v but connects via cathode biasing (as per Rod's instruction sheet).
I currently have signal 0v of input sockets connected directly to 0v of ouput sockets and to star earth board. Have tried some variatioons of this connection scheme without noticeable results.

Amp chassis is mostly wooden with an alu plate for tube sockets and heatsink for BJT's on heater boards, and it is currently not connected to earth at all. PSU chassis has alu top plate, which is connected by a thick wire to 0v star point (1st capacitor 0v) and from there a short wire runs to IEC plug's earth pin.
Unfortunately, I don't have a camera and my cellphone can't download pics to my computer. (Takes lousy pics anyway).
 
OK, the connection is done and the amp is now safer.

I have a tapping in the transformer for a lower HT - 140v if I remember correctly. I had it wound that way in case I needed to resort to SS rectification. I would prefer to stick with valve rectification if possible, but it is there if nothing else works. I have been struggling with this amp for a long time and am feeling quite disheartened.
 
Hi again,

if the ground connection doesn't help, some systematic search is due. Measure the ripple in the B+ and across the filaments. PSUD is fine for theoretical analysis, but if something is wired wrongly, or a part is faulty, reality can be very different.
How is the grid conencted? To a pot? connect the grid directly to ground and check for hum.

No reason to be disheartened. This should be fun! Once you solved such a problem you will have learned a lot from that

Thomas