I have a few EL503 in a drawer that I want to try someday. The spec says: Pdiss 27W, max DC 200mA, Zp-p 2400 ohms, AB output power is 40W with 12.2 Vrms on the G1, B+ is 265v. I haven't found any data on EL503 as single ended output tube. Theoretically, I guess, it should be possible to build a simple SE amplifier with a 12AX7 followed by a EL503 and get more power and gain than a EL34 based amp, at a low 250-280v B+ with a output transformer with around 2000 ohm primary impedence (or maybe less). I've never tried because the EL503 is getting even more rare than the ELL80, and I've not found any cheap tube to replace the EL503 should they eventually fail. Classic amplifiers with EL503 are sometimes modified to use EL34 or EL36, but neither of them has the same high gain and low output impedence requirements. Is there maybe a inexpensive sweep tube in the same 27W power class that can be paired with a single stage 12AX7 / 6SL7 input triode to get a working SE amplifier, with gain to spare for a little global feedback and capable of full power at 1Vpp input? This vould be a easy and inexpensive build to try. My EL34 build with a 300V B+ does need two 6sn7 stages before the EL34 to get enough gain, and I'm not happy of the results. I tried with a single 6SL7 stage but it sounds even worse.
EL803, with a gm of 23,000 and Pdiss of 27 Watt, has only one replacement that I know of, 8417. Which is equally rare. (well EL520 too, but even more scarce) However, one could make something up by paralleling.
3x 12HL7 in parallel would offer 60,000 gm and 30 Watts Pdiss summed. They were on sale at $1 each, but normally are around $3 each. (9 pin 10 Watt frame grid tube, 20,000 gm)
Probably safer to go with 4x in parallel for more conservative Watts. (I don't really trust their 10 Watt rating with what look like 6 Watt plates!)
12GN7 is another similar type. ($6, 9 pin)
3x (or 4x) 6HB6 would work similarly too. (9 pin 10 Watt, 20,000 gm non frame grid, $3)
Or, 2x 21LG6 in parallel would offer 23,000 gm and 56 Watt Pdiss. Normally around $4 each. (12 pin with a plate cap, 28 Watt)
A single 21LG6 could sub for the EL803 using a 12HL7/12GN7/or 6HB6 high gain driver. This would seem the least trouble to implement.
Any of these tubes are actually quite decent tubes. Normally I would think of the 12HL7 or 12GN7 as a beefy driver tube. (a hyper 12BY7)
The 12HL7/12GN7 frame grids might be microphonic.
6HB6 could be an output tube or a super driver tube.
21LG6 is a conventional TV Sweep tube. (6LG6 are scarce)
Unfortunately, some Hong-Kong buyer bought up nearly the world supply (30,000) of the 6HB6 tubes recently. They likely are scarce now.
12HL7 may not be quite so plentiful either, after the $1 SALE depleted some of them.
https://frank.pocnet.net/sheets/010/e/EL503.pdf
https://frank.pocnet.net/sheets/123/6/6LG6.pdf
https://frank.pocnet.net/sheets/135/6/6HB6.pdf
https://frank.pocnet.net/sheets/135/1/12HL7.pdf
https://frank.pocnet.net/sheets/135/1/12GN7A.pdf
EL503, with g2/g1 Mu of 13, would have 13x1000000/23000 = 565 Ohm Ri as triode at 110 mA.
For SE, usually 3 to 5 times that for the SE OT primary Z. (needs to be further corrected for gm at the actual operating current)
21LG6, with g2/g1 Mu of 3.6, would have 3.6x1000000/11500 = 313 Ohm Ri as triode at 90 mA. But low V spec/operation of g2 likely precludes it's use as a triode.
..
3x 12HL7 in parallel would offer 60,000 gm and 30 Watts Pdiss summed. They were on sale at $1 each, but normally are around $3 each. (9 pin 10 Watt frame grid tube, 20,000 gm)
Probably safer to go with 4x in parallel for more conservative Watts. (I don't really trust their 10 Watt rating with what look like 6 Watt plates!)
12GN7 is another similar type. ($6, 9 pin)
3x (or 4x) 6HB6 would work similarly too. (9 pin 10 Watt, 20,000 gm non frame grid, $3)
Or, 2x 21LG6 in parallel would offer 23,000 gm and 56 Watt Pdiss. Normally around $4 each. (12 pin with a plate cap, 28 Watt)
A single 21LG6 could sub for the EL803 using a 12HL7/12GN7/or 6HB6 high gain driver. This would seem the least trouble to implement.
Any of these tubes are actually quite decent tubes. Normally I would think of the 12HL7 or 12GN7 as a beefy driver tube. (a hyper 12BY7)
The 12HL7/12GN7 frame grids might be microphonic.
6HB6 could be an output tube or a super driver tube.
21LG6 is a conventional TV Sweep tube. (6LG6 are scarce)
Unfortunately, some Hong-Kong buyer bought up nearly the world supply (30,000) of the 6HB6 tubes recently. They likely are scarce now.
12HL7 may not be quite so plentiful either, after the $1 SALE depleted some of them.
https://frank.pocnet.net/sheets/010/e/EL503.pdf
https://frank.pocnet.net/sheets/123/6/6LG6.pdf
https://frank.pocnet.net/sheets/135/6/6HB6.pdf
https://frank.pocnet.net/sheets/135/1/12HL7.pdf
https://frank.pocnet.net/sheets/135/1/12GN7A.pdf
EL503, with g2/g1 Mu of 13, would have 13x1000000/23000 = 565 Ohm Ri as triode at 110 mA.
For SE, usually 3 to 5 times that for the SE OT primary Z. (needs to be further corrected for gm at the actual operating current)
21LG6, with g2/g1 Mu of 3.6, would have 3.6x1000000/11500 = 313 Ohm Ri as triode at 90 mA. But low V spec/operation of g2 likely precludes it's use as a triode.
..
Last edited:
A little note on that P-P OT impedance formula in post #20:
Z = k SQRT(Watts rating) / (DC Max current rating)
where k is determined from analyzing a group of known similar tube cases,
and k varies over a range, as emphasis is on power versus low distortion, and also on the amount of class A overlap. (more power ---> lower k) (more overlap ---> lower k)
That's an heuristic formula based on tube designs that scale the current AND the voltage rating of tubes by the SQRT of Power rating. Which is -usually- a sensible assumption within a design class of similar tubes.
However it does NOT work well for paralleling tubes, where the voltage rating does not increase any. (also not working well, for the same k factor, for tubes like the 6CW5/EL86 where the voltage rating remains unusually low, essentially an internally paralleled small tube design) For paralleling tubes, the heuristic formula would be:
Z = j (total Watts rating)/ (total DC Max current rating)^2
(^2 is for Squared)
"constant" j is similarly determined from analyzing a group of known similar tube cases (or single pair case) and varies over a range with emphasis on power versus low distortion, and also on the amount of class A overlap.
TV Sweep tube designs are usually similarly scaled designs, so the 1st formula works well between different Sweep tubes. (except 6L6, 6V6...)
Audio tubes look like 1/2 a Sweep tube for current, but similar scaled V rating with power. So just a different k factor range will work for them using the 1st formula for single tube pairs.
6CW5/EL86 is like two EL84 in parallel, so low Z loading is expected with the low V rating. But 3000 Ohms seems extreme, since that makes the k for 6CW5 half of TV Sweeps. Should be at least 4000 Ohm loading.
6L6 at 3800 Ohms is an anomaly for Watts output advertising for sure. This tube would be pushing near its max DC current rating (110 mA max DC) continuously for full power. (audio crest factor saves it from a short life) 5000 or 6600 Ohms seems more reasonable. Its k factor should be more like similar that of TV Sweeps, not half! (its lower V rating than Sweeps, brings it down from 2x k !!)
..
Z = k SQRT(Watts rating) / (DC Max current rating)
where k is determined from analyzing a group of known similar tube cases,
and k varies over a range, as emphasis is on power versus low distortion, and also on the amount of class A overlap. (more power ---> lower k) (more overlap ---> lower k)
That's an heuristic formula based on tube designs that scale the current AND the voltage rating of tubes by the SQRT of Power rating. Which is -usually- a sensible assumption within a design class of similar tubes.
However it does NOT work well for paralleling tubes, where the voltage rating does not increase any. (also not working well, for the same k factor, for tubes like the 6CW5/EL86 where the voltage rating remains unusually low, essentially an internally paralleled small tube design) For paralleling tubes, the heuristic formula would be:
Z = j (total Watts rating)/ (total DC Max current rating)^2
(^2 is for Squared)
"constant" j is similarly determined from analyzing a group of known similar tube cases (or single pair case) and varies over a range with emphasis on power versus low distortion, and also on the amount of class A overlap.
TV Sweep tube designs are usually similarly scaled designs, so the 1st formula works well between different Sweep tubes. (except 6L6, 6V6...)
Audio tubes look like 1/2 a Sweep tube for current, but similar scaled V rating with power. So just a different k factor range will work for them using the 1st formula for single tube pairs.
6CW5/EL86 is like two EL84 in parallel, so low Z loading is expected with the low V rating. But 3000 Ohms seems extreme, since that makes the k for 6CW5 half of TV Sweeps. Should be at least 4000 Ohm loading.
6L6 at 3800 Ohms is an anomaly for Watts output advertising for sure. This tube would be pushing near its max DC current rating (110 mA max DC) continuously for full power. (audio crest factor saves it from a short life) 5000 or 6600 Ohms seems more reasonable. Its k factor should be more like similar that of TV Sweeps, not half! (its lower V rating than Sweeps, brings it down from 2x k !!)
..
Last edited:
I should also mention that the actual Z primary used in a design depends on the actual B+ used. These heuristic formulas above are just intended to estimate where the Zpri (and associated B+) should -best- be based on similar successful designs.
Thank for the advice, good to know that there are several options to check. The 21LG6 seems to be a interesting sweep tube, worth exploring further, I didn't knew it.
I will check further the suggestion of paralleling 3-4 12HL7 tubes, but I see a potential issue. Frame grid tubes characteristics have a reputation to be anything but constant from tube to tube. Connecting them in parallel may be unpractical without a tube tester. These TV tubes don't seems to be sold as matched pairs or quartets. But maybe this word of mouth is not true.
I also have a not so good experience with frame grid TV pentodes. I tried to build a SPUD amp with the high gm frame grid tube EL802 some times ago, but I was unable to stop HF oscillations. After a while I gave up and exchanged it for a 6E5P, bought as matched pair for a low price. This worked flawlessly from the beginning and does sounds pretty good. The 6HB6 is still available, fortunately. I would check it first.
Glad to see that my back of the envelope calculation for EL503 SE OPT in triode mode does match with your experience, 2K is right in the middle of your suggested range. Would 3-3.5k be OK for a EL503 pentode configuration? This would basically be the average, easy to source and reuse KT88 SE output transformer.
I will check further the suggestion of paralleling 3-4 12HL7 tubes, but I see a potential issue. Frame grid tubes characteristics have a reputation to be anything but constant from tube to tube. Connecting them in parallel may be unpractical without a tube tester. These TV tubes don't seems to be sold as matched pairs or quartets. But maybe this word of mouth is not true.
I also have a not so good experience with frame grid TV pentodes. I tried to build a SPUD amp with the high gm frame grid tube EL802 some times ago, but I was unable to stop HF oscillations. After a while I gave up and exchanged it for a 6E5P, bought as matched pair for a low price. This worked flawlessly from the beginning and does sounds pretty good. The 6HB6 is still available, fortunately. I would check it first.
Glad to see that my back of the envelope calculation for EL503 SE OPT in triode mode does match with your experience, 2K is right in the middle of your suggested range. Would 3-3.5k be OK for a EL503 pentode configuration? This would basically be the average, easy to source and reuse KT88 SE output transformer.
I just came across another very suitable substitute for the EL803. The E130L.
Would have a lower Ri in triode too. (g2/g1 Mu 6.5 ) Seem to be around $25 on Ebay.
Datasheet gives triode curves.
https://frank.pocnet.net/sheets/009/e/E130L.pdf
e130l tube | eBay
Would have a lower Ri in triode too. (g2/g1 Mu 6.5 ) Seem to be around $25 on Ebay.
Datasheet gives triode curves.
https://frank.pocnet.net/sheets/009/e/E130L.pdf
e130l tube | eBay
Last edited:
With E130L the problem is very low the maximum permissible voltage Ug2 (max ~ 260-270V!).
Through the practice of working with the EL503, I have found that it is well tolerated by the voltage Ug2 to ~ 350-360V.
The electronics F2a and F2a11 have a somewhat lower mutual conductance, well tolerance Ug2~450V, but their price is very high!
Through the practice of working with the EL503, I have found that it is well tolerated by the voltage Ug2 to ~ 350-360V.
The electronics F2a and F2a11 have a somewhat lower mutual conductance, well tolerance Ug2~450V, but their price is very high!
I'm surprised they would give triode curves on the E130L datasheet then. Certainly TV Sweep tubes with similar 250 V screen grid max's are problematic in triode though.
One thing I notice about this tube is the low screen current compared to plate current. The frame grid tubes (like 12HL7 or 12GN7) are like 45% of plate current at the knee. While the E130L is around 10% at the knee. So I'm guessing the E130L is an aligned grid tube to get low screen current.
One thing I notice about this tube is the low screen current compared to plate current. The frame grid tubes (like 12HL7 or 12GN7) are like 45% of plate current at the knee. While the E130L is around 10% at the knee. So I'm guessing the E130L is an aligned grid tube to get low screen current.
Just as a matter of interest, the QQE03-12 makes for a very sweet sounding compact PP amp. Can be configured as a self splitter.
Shoog
Shoog
I looked at the E130L datasheet triode curves again. Looks like you just have to run with 120V to 150V B+ and all is well for max ratings. It's 300 mA max DC rating will compensate for the low B+ with a low Zprimary OT.
I also ordered an E130L off Ebay to evaluate. I'm surprised that the plate size is about like a 17.5 Watt TV Sweep tube. The big grid support posts make it apparent that it has frame grids for g1 and g2. Although the grid wire spacing is just a little closer than a 21HB5 (the closest grid wire spacing TV Sweep I've seen).
So the E130L 10% screen current (of plate current at max knee) is likely from grid wire alignment.
Looking at the curves for an 8417 tube, I see screen current at 18.5% of plate current at max knee.
Looking at EL3010 I see screen current at about 25%
Looking at EL503 I see screen current at 10%
for 21LG6 I see 5%
And for 21HB5 I see screen current at 3% to 5%
Low screen current might indicate good possibilities for triode operation. However 21LG6 only has a 200V max g2 rating and 21HB5 only a 220V max g2 rating. Both have plenty of current capability to run at low B+ like the E130L at least.
21HB5 have no plate cap. Two of them in parallel would give an effective E130L or EL503. They used to be real cheap back a few years ($1), but are now around $6. 17/22KV6 would be another similar possibility, $6 too.
21HB5 using Crazy/Twin drive below:
..
I also ordered an E130L off Ebay to evaluate. I'm surprised that the plate size is about like a 17.5 Watt TV Sweep tube. The big grid support posts make it apparent that it has frame grids for g1 and g2. Although the grid wire spacing is just a little closer than a 21HB5 (the closest grid wire spacing TV Sweep I've seen).
So the E130L 10% screen current (of plate current at max knee) is likely from grid wire alignment.
Looking at the curves for an 8417 tube, I see screen current at 18.5% of plate current at max knee.
Looking at EL3010 I see screen current at about 25%
Looking at EL503 I see screen current at 10%
for 21LG6 I see 5%
And for 21HB5 I see screen current at 3% to 5%
Low screen current might indicate good possibilities for triode operation. However 21LG6 only has a 200V max g2 rating and 21HB5 only a 220V max g2 rating. Both have plenty of current capability to run at low B+ like the E130L at least.
21HB5 have no plate cap. Two of them in parallel would give an effective E130L or EL503. They used to be real cheap back a few years ($1), but are now around $6. 17/22KV6 would be another similar possibility, $6 too.
21HB5 using Crazy/Twin drive below:
..
Attachments
Last edited:
Take some 50/80 Watt amplifier design like below (linked) and sub in 6X paralleled ELL80 tubes, or 2X (or 3X) paralleled 6CW5/EL86 tubes, (6CW5 1/30 cost of the ELL80 setup for the same sound) for each of the 31LQ6 tubes shown.
Some adjustments then needed for B+ V and OT primary Z. Screen V too maybe. You will want to check the grid input capacitance for the paralleled ELL80, 6CW5 versus the 6LQ6 tubes.
The 100K resistors with the arrows connect to -Bias V adjustor pots. You will likely want individual ones for each tube/grid section. (not likely going to find matched tubes for ELL80) Check maximum grid bias resistor values from datasheet for each case (and then paralleled). Compare with 6LQ6 case. (6DJ8 drivers have to drive all these bias resistors and grid input capacitances.)
Some variable HV power supplies would be helpful for setting things up, but one could go for the final power supply/xfmr too. (have to figure AC voltages and current requirements for B+, heater power, bias winding)
http://webpages.charter.net/dawill/tmoranwms/Circuits/31LQ6_PP.gif
Philips ECG NOS 6CW5 EL86 Red Graphics Vacuum Tubes Matched and Strong! (5) | eBay
http://www.edcorusa.com/cxppseries#/specFilters=9m!#-!264!##!430
http://tubedata.milbert.com/sheets/135/6/6LQ6.pdf
http://tubedata.milbert.com/sheets/093/6/6CW5.pdf
http://tubedata.milbert.com/sheets/124/e/ELL80.pdf
.
Some adjustments then needed for B+ V and OT primary Z. Screen V too maybe. You will want to check the grid input capacitance for the paralleled ELL80, 6CW5 versus the 6LQ6 tubes.
The 100K resistors with the arrows connect to -Bias V adjustor pots. You will likely want individual ones for each tube/grid section. (not likely going to find matched tubes for ELL80) Check maximum grid bias resistor values from datasheet for each case (and then paralleled). Compare with 6LQ6 case. (6DJ8 drivers have to drive all these bias resistors and grid input capacitances.)
Some variable HV power supplies would be helpful for setting things up, but one could go for the final power supply/xfmr too. (have to figure AC voltages and current requirements for B+, heater power, bias winding)
http://webpages.charter.net/dawill/tmoranwms/Circuits/31LQ6_PP.gif
Philips ECG NOS 6CW5 EL86 Red Graphics Vacuum Tubes Matched and Strong! (5) | eBay
http://www.edcorusa.com/cxppseries#/specFilters=9m!#-!264!##!430
http://tubedata.milbert.com/sheets/135/6/6LQ6.pdf
http://tubedata.milbert.com/sheets/093/6/6CW5.pdf
http://tubedata.milbert.com/sheets/124/e/ELL80.pdf
.
Last edited:
Oh, and sum up the paralleled output tube gm's to compare with the 6LQ6 case. That probably will exceed the 6LQ6 gm, so no additional drive signal swing would be needed then.
Figure heater current for all the paralleled tubes + the front end tubes.
With an adjustable power supply or adjustable regulator for the screen V, you could try out most any TV Sweep tubes in the amplifier too. Might be fun to compare (TV beamer tubes) with the paralleled pentode tubes for sound.
Figure heater current for all the paralleled tubes + the front end tubes.
With an adjustable power supply or adjustable regulator for the screen V, you could try out most any TV Sweep tubes in the amplifier too. Might be fun to compare (TV beamer tubes) with the paralleled pentode tubes for sound.
Last edited:
Another (60 Watt) OT option.
EDCOR - CXPP60-2.4K
With multiple secondary impedance outputs on the OT, one could put say an 8 Ohm speaker on the 4 Ohm output and run with half the paralleled tubes. (then "only" $240 of ELL80 tubes or $12 of 6CW5/EL86 ) Can try the amp before fully committing to all those tubes that way.
And get yourself a couple of $4 6HJ5 or 21LG6 or 31LQ6 tubes to try too, 1st (before buying all those ELL80 tubes!). You might decide that's pretty darn good already.
EDCOR - CXPP60-2.4K
With multiple secondary impedance outputs on the OT, one could put say an 8 Ohm speaker on the 4 Ohm output and run with half the paralleled tubes. (then "only" $240 of ELL80 tubes or $12 of 6CW5/EL86 ) Can try the amp before fully committing to all those tubes that way.
And get yourself a couple of $4 6HJ5 or 21LG6 or 31LQ6 tubes to try too, 1st (before buying all those ELL80 tubes!). You might decide that's pretty darn good already.
Last edited:
- Status
- Not open for further replies.