• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

50W monoblock "Engineers Amp"

Huhh.
Usually the secondary is wound for a higher voltage than the nameplate V, to allow for resistance losses. So they should run cooler backwards with reduced magnetizing current. But some of those industrial xfmrs are just plain over economized on steel. I've got a Hammond 500 VA industrial xfmr here ($15 Mendelsons Elect.) that weighs half as much as the Hammond 378CX tube power xfmr (465 VA) ($170 now). The industrial xfmr draws 3X the magnetizing current too, and is marked with an 80 degree C temp rise. It is quiet though. Laminations are strip welded. Most of the temp. rise is likely from copper losses at full load though. The industrials skimp on copper too. You just get high temp insulation.

On the other hand, I have a GE 385 VA isolated 55VAC Volt-Pac Variac (from Fair-Radio) that runs with 1/8 the magnetizing current of the Hammond 378CX. I thought I missed a connection when I tested it on the scope with a current probe. I'm sure it's continuous M6 steel (and toroid) to do that.

.
 
Last edited:
Anyone able to comment on how much buzz/hum from the Hammond power xfmr 290FX versus the 291FX? I know the 300 series (373FX) are dead quiet, but they cost almost 2X.

I haven't tried the 291FX, but at least on my line voltage (120V 60Hz, pretty clean) in this amp the 290FX is dead quiet. If you have any DC or bad distortion on your power you may have more noise, though.

Pete
 
Updates

I optimized the NFB network for the Hammond OPT, and added a phase compensation cap: C1 33pF. BOMs and schematic have been updated.

I also made a bunch of measurements. They are posted on the web page. Here's a teaser:

DCPP_MB_THD_Power.jpg
 
Last edited:
Another 7 pin, high gm tube that will fit the 6CB6/6GU5 socket (pin compatible) is the 6JK6. However these are expensive, and the boxes are often marked 6JK6/6JL6. 6JL6 is a remote cut-off version, so not so hot (who knows what you are actually getting).

But, 5JK6 are available cheaply ($3 each or $0.75/10 ) Would need some SS diodes in series with the heater to drop the 6.3V slightly. Or run it off the 5 VAC winding on the power xfmr if available. (the Hammond 373FX power xfmr has 5V and 6.3V)



https://frank.pocnet.net/sheets/201/6/6JK6.pdf gm = 18000 (frame grid)

https://frank.pocnet.net/sheets/093/6/6EW6.pdf gm = 14000

https://frank.pocnet.net/sheets/093/6/6CB6A.pdf gm = 8000

https://frank.pocnet.net/sheets/049/6/6GU5.pdf gm = 15500 (remote cut-off)


Or could make a 9 Pin to 7 Pin adapter for 6JC6 (gm = 16000, frame grid, $1):

https://frank.pocnet.net/sheets/135/6/6JC6A.pdf



..
 
Last edited:
Maybe that's why the 6J11 datasheet doesn't give any curves. Your supposed to know its a 6EW6. 6BN11 definitely looks just a little different than the 6EW6. (improved maybe?) The JEDEC sheets say its a minor mod to the 6J11.

I think they made some tweeks to many of the original tubes when they put them onto 12 pins. Interesting to see 21HJ5 (& 6HJ5) listed as a 6DQ5. The g2/g1 Mu factor is a little different however. (4.2 and 3.3) 6HF5 is listed as a 6DQ5 also, but Mu is 3.0. Of course those do depend on the operating point chosen. And maybe some measurements were more accurate than 10 years earlier.
 
Last edited:
And maybe some measurements were more accurate than 10 years earlier

GE and others kept improving the old tubes during their production run, especially the horizontal sweep tube. It was the #1 failure item in TV sets from the 50's through to the solid state era. Vertical sweep and damper tubes made the top 10 list as did most electrolytic caps and the B+ rectifier tube while it still existed.

For a study in VT evolution look at the early sweeps, like the 6BQ6 and 6AV5's. How many different flavors can you find, arrange them chronologically, they don't get smaller with time.

The tube manufacturers also learned from Ford, make one car, available in any color you want as long as it was black. GE especially, learned early that making several dozen different grids and plates was expensive, and error prone. Somewhere in the 60's most sweeps got the "corporate 18 watt plate" even some of the 12 watt tubes. I'm guessing that they did the same thing with grids.

We (Motorola) didn't learn that lesson until the early 80's. We had 20+ different kit numbers for minor tweaks on the same PC board. Two way radios for police are made to order. Each batch is different. Tracking and making sure the right stuff went into each radio was a nightmare. Today the boards are all the same, they just get the right software loaded into them near the end of the product flow.

Note, there is a popular Chinese oscilloscope that has taken this a bit too far. The entire line of scopes use the same hardware. You can buy the $800 scope, and reflash it so it becomes the $2200 scope.
 
Here are some curve tracings for 6EW6, 6J11, and 6BN11

All are at 2 mA/div Vert.
50V/div Horiz.
0.08 Vsteps on grid1

1) 6EW6 Vg2 = 133 V, RCA
2) 6J11 Vg2 = 96 V, SYL
3) 6BN11 Vg2 = 125 V, RCA

6J11 and 6BN11 looked identical inside, probably made in the same place.

I adjusted Vg2 for each of them to get the same max curve. So the supposed offset in bias for the 6J11 would be hidden. Using Mu = 70 for the 6BN11 or 6J11, a 0.5V offset in Vg1 would show up as 35V difference in Vg2. Which is just about what I see, 96V and 125V.

6J11 and 6BN11 look the about same to me, other than that offset. Hard to get identical tubes to match this well! Close match for both to the 6EW6 too, except 6BN11 agrees best with the 6EW6 for Vg2, I was expecting the 6J11 to agree the best. So I think maybe that equivalence table should say the 6BN11 is equivalent to the 6EW6, and the 6J11 has the offset modification. ???

..
 

Attachments

  • rsz_6ew6_p08vs_133vg2_2ma_50v.jpg
    rsz_6ew6_p08vs_133vg2_2ma_50v.jpg
    102.7 KB · Views: 1,080
  • rsz_6j11_p08vs_96v2_2ma_50v.jpg
    rsz_6j11_p08vs_96v2_2ma_50v.jpg
    100 KB · Views: 1,069
  • rsz_6bn11_p08vs_125v2_2ma_50v.jpg
    rsz_6bn11_p08vs_125v2_2ma_50v.jpg
    99.7 KB · Views: 1,070
Last edited:
Feedback compensation and IMD

After more testing and listening... I discovered an HF oscillation when driving certain speakers. Back to the bench...

It seems the 33pF compensation cap was making the amp unstable with an open load. Same happened with a speaker that had very high impedance at >30kHz (which many do).

The oscillation could be cured with a 10 ohm load resistor across the output. But I opted to remove the compensation cap. So the NFB is now again just a resistor divider. This causes a little overshoot when excited by a square wave.

For good, or for bad, depending on your philosophy - that makes the HF response extend to >100kHz. So the amp is flat from about 5Hz to 100kHz.

I also did some quick IMD measurements. SMPTE IMD @ 1W out is about 0.01%. At 10W, it rises to about 0.2%.

Hera are some scope shots showing square wave (5kHz) response and rise/fall with the comp cap removed. Note the time scales. Slew rate is > 1V/uS. More over/undershoot than I would like, but also the input square wave (yellow trace) is very high bandwidth, higher than any normal audio source...

Pete
 

Attachments

  • DCPP_MB_5khz_sq.jpg
    DCPP_MB_5khz_sq.jpg
    53.7 KB · Views: 1,054
  • DCPP_MB_5khz_rise.jpg
    DCPP_MB_5khz_rise.jpg
    58.3 KB · Views: 1,032
  • DCPP_MB_5khz_fall.jpg
    DCPP_MB_5khz_fall.jpg
    58.1 KB · Views: 139
Correction to post #52!

In post #52 I mentioned the 6LT8 as being like 1/2 a 6BN11 etc, and it does look that way on paper (datasheet). But after testing two (GE and Westinghouse) on the curve tracer, versus the 6BN11, they are NO where near the same. 6LT8 has very curved over knees and from 20% to 40% less gm than a 6BN11. Junk bin those 6LT8s. Unless I just got two bad ones. The datasheet does show sharp knees.

https://frank.pocnet.net/sheets/135/6/6LT8.pdf
https://frank.pocnet.net/sheets/123/6/6BN11.pdf
 
Last edited:
6HJ5 tubes

So... it seems that hundreds of 6HJ5 tubes have disappeared from tube vendors stocks in the last week or two. They are getting hard to find now.

I guess people have been buying them speculatively in large batches. Kind of a shame, if you ask me.

I hope to secure a supply of them myself, so I can get you tubes if you buy a board. Stay tuned.

Pete