Hi,
I need some input....I was gifted a Luxman MQ68C a while ago. Originally, the power tubes are the obsolete 50CA10 in triode mode push pull fixed bias with a plate voltage of 400VDC and a 6K OPT. The tube is capable of 35W in that way, Luxman rates the amp at 30W. The original tubes are long gone, and the Chinese replacements have a dubious reputation and are not inexpensive.
I have been putting this project off because I am not happy with the idea of running EL34 in this amp. I did install a 6A 6.3VAC transformer to power the new power tube heaters, and installed octal sockets. The PT has a 0-360V 280mA HV winding on the PT, the 50V heater will be released from duty, reducing stress on the PT.
The original EL34 would put out 19W according to the datasheet, and new production tubes would obviously be less.
The alternative is 6550 in triode PP, this does around 25W, still not at the 35W mark for the 50CA10.
The KT88 does not fit, the sides of the bottles touch. *yikes*
So then I looked at alternatives and came up with two options that I am considering:
1) KT90 biased at 50-60 mA in triode mode. The skinny bottles fit well in the amp.
2) JJ EL509S in Enhanced Triode mode (screen driven). This way, the bias current will be about 10mA per tube and this will make the PT run much cooler. I may not have enough drive from the inverter to get enough power from this tube. But that can be fixed with a Mosfet.
Has anyone done this before, or anything similar? What are your thoughts on these approaches?
I need some input....I was gifted a Luxman MQ68C a while ago. Originally, the power tubes are the obsolete 50CA10 in triode mode push pull fixed bias with a plate voltage of 400VDC and a 6K OPT. The tube is capable of 35W in that way, Luxman rates the amp at 30W. The original tubes are long gone, and the Chinese replacements have a dubious reputation and are not inexpensive.
I have been putting this project off because I am not happy with the idea of running EL34 in this amp. I did install a 6A 6.3VAC transformer to power the new power tube heaters, and installed octal sockets. The PT has a 0-360V 280mA HV winding on the PT, the 50V heater will be released from duty, reducing stress on the PT.
The original EL34 would put out 19W according to the datasheet, and new production tubes would obviously be less.
The alternative is 6550 in triode PP, this does around 25W, still not at the 35W mark for the 50CA10.
The KT88 does not fit, the sides of the bottles touch. *yikes*
So then I looked at alternatives and came up with two options that I am considering:
1) KT90 biased at 50-60 mA in triode mode. The skinny bottles fit well in the amp.
2) JJ EL509S in Enhanced Triode mode (screen driven). This way, the bias current will be about 10mA per tube and this will make the PT run much cooler. I may not have enough drive from the inverter to get enough power from this tube. But that can be fixed with a Mosfet.
Has anyone done this before, or anything similar? What are your thoughts on these approaches?
Another option is the 7591 in full pentode mode, with regulated g2 B+. The 7591 draws less heater current than the EL34 and is as easy to drive as an EL84. Sufficient open loop gain to support the requisite NFB is easy to achieve.
BTW, triode wire KT88s, even if they fit, will get you no more power than GE 6550As.
BTW, triode wire KT88s, even if they fit, will get you no more power than GE 6550As.
Thanks for the reply fellow Jersian! I am in the Morristown area myself. 7591....not a bad idea at all, the 7591 is very nice tube, but I want to stay with the triode sound that the unit was designed with.
True on the GE 6550A tubes. The thing is that I do not feel like digging into my GE 6550A stash for this Hifi amp. I noticed in the guitar amps I make that the new production 6550As are less powerful than new production KT88. But that is push pull pentode mode, triode may be different.
Also, I keep reading that 6550s wear out quickly in triode mode, much faster than KT90.
True on the GE 6550A tubes. The thing is that I do not feel like digging into my GE 6550A stash for this Hifi amp. I noticed in the guitar amps I make that the new production 6550As are less powerful than new production KT88. But that is push pull pentode mode, triode may be different.
Also, I keep reading that 6550s wear out quickly in triode mode, much faster than KT90.
OK, triode it is. Triode wired 6L6GCs in Class "AB2" will get you the desired power. As always, TANSTAAFL applies and driver circuitry complexity is the price you'll pay. George Anderson, aka Tubelab, and Pete Millett show the way with MOSFETs. Look here and here.
As far as the KT90 is concerned, good specimens (EI Type 2 and early production EI Type 3) are unobtainium.
The EH KT90 was initially very mediocre, but it has gotten better; unfortunately, (IMO) not enough better.
As far as the KT90 is concerned, good specimens (EI Type 2 and early production EI Type 3) are unobtainium.
The EH KT90 was initially very mediocre, but it has gotten better; unfortunately, (IMO) not enough better.
FWIW, i have a pair of 50AC10's and a quad of 6AC10's, new production Shouguang Care about U$75 each...
i repaired a Luxman LX-38, i converted to cathode resistor bias.....
couldn't tell the difference...the amp sounds good...
i repaired a Luxman LX-38, i converted to cathode resistor bias.....
couldn't tell the difference...the amp sounds good...
OK, triode it is. Triode wired 6L6GCs in Class "AB2" will get you the desired power. As always, TANSTAAFL applies and driver circuitry complexity is the price you'll pay. George Anderson, aka Tubelab, and Pete Millett show the way with MOSFETs. Look here and here.
As far as the KT90 is concerned, good specimens (EI Type 2 and early production EI Type 3) are unobtainium.
Thanks for the reply. Driver complexity, I see your point. I am routinely doing cathode follower/mosfet driver stages in my shop to drive 6L6GC and 6550 push pull pentode power amps.
For this Luxman project, I was hoping we could find a tube that is similar in performance to the old 50CA10, and at least for now, that we could not deal with the additional complexities till I find a reason to warrant such complexities. heh. The reality is that space is tight in the Luxman chassis. (I promise that I will find flaws with the design, and I will hate myself for wanting a fix.)
Are there other tubes that are worth looking into?
my MQ88 with just half inch wider case perfectly runs on KT-88. Though, yours for that will require a new mains trafo for 6.3V filament + replacing the sockets, quite a bit of work anyway...
Yea no. I fitted KT88 in the octal sockets that I put in. The KT88s were slanted in the sockets and still touching...bad. There has to be a better solution.
they look like tv scanning tube internally connected as triodes...
screen drive? the ef86/6aq8 tubes may not have enough juice...
I see. Well, in this case I'd go for either compactron-based 6LU8 or the novar 6LR8 ones, still many of these around. This could be somewhat implementation of Baby Huey - the first cascades could be built on them instead. I think these would fit perfectly - they even look alike the original 50CA10.
The 6LU8 and the 6LR8 are not powerful enough for this task. I am looking to get 30W triode without having to drive the snot out of the power tubes. It would be nice if they would last a bit. 🙂 BTW, I am looking to drive ADS L1230 speakers.
Yes, it is a bit of work.
I was gifted the MQ68C because it was partially stripped by someone who thought it could be used as a keyboard amp. So they attempted to mod the preamp/driver section of one channel. After that, it changed hands several times, and ended up with me for pretty much a full rebuild. One channel is intact, the other one was stripped. I have since rebuilt that channel.
I added a 6.3VAC 6A transformer, and put octal sockets in, with the idea of using EL34 or 6550. But after looking at the tube data, and speaking with other Lux owners who have heard the EL34 and 6550 modified MQ68C amps next to the originals, I am not so sure if I want to use those tubes. So I am looking for a more readily available octal tube that performs closer to the 50CA10.
Push pull pair TRIODE
Plate voltage 400V
Grid -43V
Input signal 60V RMS
Bias current at idle 100mA
Plate current at max signal 150mA
Plate load 5000 Ohms
Output power 34W
Distortion 2.50%
IMO, 2.5% THD at full power O/P is unacceptable, as it is highly audible. For me, the target is 0.3 % THD, which is below the audibility threshold. Claims for smaller numbers in either tube or SS equipment is just marketing "hype".
FWIW, I think you have to bite the 6L6GC/"AB2" bullet.
FWIW, I think you have to bite the 6L6GC/"AB2" bullet.
Thank you for that. I will certainly try the 6L6GC. The amp will get IRF820 source follower drivers after the PI, so it will be able to push the power tubes into plate saturation.
The IRFBC20 is an improvement over the IRF820. Capacitances, especially the critical reverse transfer capacitance, are smaller.
As shown by George and Pete, DC couple the FET to the O/P tube's g1. You have to drag g1 positive (Class "AB2"), to obtain full power O/P.
You can consider taking a page out of Stu Hegeman's book and nest NFB loops, as he did in the H/K Cit. 2. Connect the O/P tube plates to the phase splitter grids. Holding the amount of NFB in the outer, global, loop to a small amount is a very good thing.
As shown by George and Pete, DC couple the FET to the O/P tube's g1. You have to drag g1 positive (Class "AB2"), to obtain full power O/P.
You can consider taking a page out of Stu Hegeman's book and nest NFB loops, as he did in the H/K Cit. 2. Connect the O/P tube plates to the phase splitter grids. Holding the amount of NFB in the outer, global, loop to a small amount is a very good thing.
100mA idle? A bit unnecessarily unreasonable IMO. I run KT-88s on 35mA without any sign of crossover dist. A pair of LU/LR will give you the very decent 20+ Wa, I can't think why I would need more for everyday listening. In half power the amp shakes the chair )) by driving the very-difficult-for-tube-amp-to-drive CM9 B&W - they have the nasty impedance drop ay LFs.
in fact LR/LU would give better linearity, I was considering modding rather to these )) but achieved the result by other means, mainly by careful PSU re-design + adding the MOSFET choke, and by lowering gain + removing any FB.
To my ear now it is noticeably better than 50CA10, airy & smooth, though I did not compare them side by side. But when I visit a guy with all the original MQ68 - I do not like his sound as much as I did (and which made me buying mine in the first place). That guy does not wanna come to listen to mine... the blasphemy-modded one ))
My latest plans are to mod it to the balanced input + replace the electrolytics with same-sized film start caps + couple minor things, maybe in next year.
in fact LR/LU would give better linearity, I was considering modding rather to these )) but achieved the result by other means, mainly by careful PSU re-design + adding the MOSFET choke, and by lowering gain + removing any FB.
To my ear now it is noticeably better than 50CA10, airy & smooth, though I did not compare them side by side. But when I visit a guy with all the original MQ68 - I do not like his sound as much as I did (and which made me buying mine in the first place). That guy does not wanna come to listen to mine... the blasphemy-modded one ))
My latest plans are to mod it to the balanced input + replace the electrolytics with same-sized film start caps + couple minor things, maybe in next year.
6LU8 is a good sounding tube, i made several of those....http://www.diyaudio.com/forums/tubes-valves/183193-tonys-6lu8-push-pull-power-amp-build.html