Matching 11MS8s
So, I finally got in a batch of sixteen 11MS8 tubes and have been running them through a Hickok tester. Man, there's quite a wide variation in the tubes. I mean, this is stuff we already knew, but, wow. But, anyway, Of the sixteen, I have this continuum from which I can create seven pairs of tubes matched within about 100 micromhos of one another on the pentode. (Given the circuit topology, am I correct in thinking triode-section matching the 11MS8s is not particularly crucial for the K-12M?)
So far, I've seen two nearly-identical varieties of 11MS8s. Matsushita and Sylvania. These appear identical save that the Matsushita branded tubes in my possession have a slightly thicker support of the metal ring below the getter at the top. My Hickok doesn't seem to favor either brand. I've got mixed-brand pairs in my K-12M with no adverse affect. But, yeah, matching has definitely eliminated the glowing plates! S-5 really really needs to match the tubes they're selling. For the five minutes it takes, I'm sure it'll make a huge difference on the amp's life and the newbie builder's impression of tube amps' reliability.
If you'd like NOS matched-pentode 11MS8 pairs, contact me.
So, I finally got in a batch of sixteen 11MS8 tubes and have been running them through a Hickok tester. Man, there's quite a wide variation in the tubes. I mean, this is stuff we already knew, but, wow. But, anyway, Of the sixteen, I have this continuum from which I can create seven pairs of tubes matched within about 100 micromhos of one another on the pentode. (Given the circuit topology, am I correct in thinking triode-section matching the 11MS8s is not particularly crucial for the K-12M?)
So far, I've seen two nearly-identical varieties of 11MS8s. Matsushita and Sylvania. These appear identical save that the Matsushita branded tubes in my possession have a slightly thicker support of the metal ring below the getter at the top. My Hickok doesn't seem to favor either brand. I've got mixed-brand pairs in my K-12M with no adverse affect. But, yeah, matching has definitely eliminated the glowing plates! S-5 really really needs to match the tubes they're selling. For the five minutes it takes, I'm sure it'll make a huge difference on the amp's life and the newbie builder's impression of tube amps' reliability.
If you'd like NOS matched-pentode 11MS8 pairs, contact me.
Does tube matching make sense ?
Chromal, I dont intend to discount your efforts by any means. It is better to use matched tubes and then no matching at all but these are just my thoughts on the subject..
Does tube matching really make sense ? There are other parameters - only two that I can think of right now - that affect the matching.
Although the tubes are matched before they are used, there is no guarantee that they will age at the same rate once they are in regular duty, hence at some point down the line they will mismatch again. What would you do then ? Especially when they are driven at or over their rating all the time!
I believe P-P transformers typically have different DCR on both primary windings. How would you account for that ?
What I am basically trying to say is that tubes that match perfectly on the tube tester, are not so forever, and might no longer be perfect matches in the circuit even on day one!
Secondly, what if one tube breaks down ? You need to find out one tube that matches the working tube or change the whole set!
To me it only makes sense to measure in-circuit voltage across the cathode resistor and adjust its value to achieve a dynamic match.
Especially for the S5 K-12M kit as it uses not-so-easy-to-find-in-bulk-quantities 11MS8 and to keep costs low (remember its a budget kit) I think S5 should modify the circuit to put adjustable bias resistors. That might be more cost and time efficient than matching tubes.
Just my opinion...
Chromal, I dont intend to discount your efforts by any means. It is better to use matched tubes and then no matching at all but these are just my thoughts on the subject..
Does tube matching really make sense ? There are other parameters - only two that I can think of right now - that affect the matching.
Although the tubes are matched before they are used, there is no guarantee that they will age at the same rate once they are in regular duty, hence at some point down the line they will mismatch again. What would you do then ? Especially when they are driven at or over their rating all the time!
I believe P-P transformers typically have different DCR on both primary windings. How would you account for that ?
What I am basically trying to say is that tubes that match perfectly on the tube tester, are not so forever, and might no longer be perfect matches in the circuit even on day one!
Secondly, what if one tube breaks down ? You need to find out one tube that matches the working tube or change the whole set!
To me it only makes sense to measure in-circuit voltage across the cathode resistor and adjust its value to achieve a dynamic match.
Especially for the S5 K-12M kit as it uses not-so-easy-to-find-in-bulk-quantities 11MS8 and to keep costs low (remember its a budget kit) I think S5 should modify the circuit to put adjustable bias resistors. That might be more cost and time efficient than matching tubes.
Just my opinion...
Re: Does tube matching make sense ?
Ah, it's good of you to pointing out pitfalls here.
Though I hadn't considered this "solution" final, I feel a little more motivated to proceed with the circuit modifications I had been considering...
Hmm. Well, I think my main goal was to come up with paired tubes that weren't so wildly mismatched that they'd destroy one another quickly with a glowing plate! But, yeah, your point that this is not a final solution to a circuit design problem is well-made.
Why do manufacturers bother with matched pairs/quads at all? Is there more going into their matching than simple micromho comparisons?
If I lost a matched pair because a tube failed or changed drastically, wouldn't I simply drop another pair in, re-measure the removed tube, and add it to the spare pool for possible future utility?
Even if we set the operating point such that each tube is drawing the same current on zero-signal, what's to assure that the current flow through each tube will match over the range of the control grid voltage swing?
In the exisitng circuit, measuring the voltage across the pair's shared cathode resister doesn't say much about the current through a specific tube...
Perhaps a socketed resister arrangement or wirewound pot should be placed between each cathode and ground. I've been thinking about this ever since I assembled the kit. Adjusting the resistance between each cathode and ground seems like it would be better. It would have been nice if S-5 had gone that route in their design, though providing a $10 wirewound pot for each tube would have added 30% to the price of the kit!
One of the things that's really sort of giving me pause in all this is my general lack of knowledge about the tube. The info out there seems minimal at best, and I have yet to come across something as simple as a grid/plate curve chart.
Ah, it's good of you to pointing out pitfalls here.
Though I hadn't considered this "solution" final, I feel a little more motivated to proceed with the circuit modifications I had been considering...percy said:
Hmm. Well, I think my main goal was to come up with paired tubes that weren't so wildly mismatched that they'd destroy one another quickly with a glowing plate! But, yeah, your point that this is not a final solution to a circuit design problem is well-made.
Why do manufacturers bother with matched pairs/quads at all? Is there more going into their matching than simple micromho comparisons?
If I lost a matched pair because a tube failed or changed drastically, wouldn't I simply drop another pair in, re-measure the removed tube, and add it to the spare pool for possible future utility?
Even if we set the operating point such that each tube is drawing the same current on zero-signal, what's to assure that the current flow through each tube will match over the range of the control grid voltage swing?
In the exisitng circuit, measuring the voltage across the pair's shared cathode resister doesn't say much about the current through a specific tube...
Perhaps a socketed resister arrangement or wirewound pot should be placed between each cathode and ground. I've been thinking about this ever since I assembled the kit. Adjusting the resistance between each cathode and ground seems like it would be better. It would have been nice if S-5 had gone that route in their design, though providing a $10 wirewound pot for each tube would have added 30% to the price of the kit!
One of the things that's really sort of giving me pause in all this is my general lack of knowledge about the tube. The info out there seems minimal at best, and I have yet to come across something as simple as a grid/plate curve chart.
Re: Re: Does tube matching make sense ?
or in the seperate pot/resistor arrangement, you could just change one tube and save half the cost and might still be better off than using a "matched pair". Not to mention vendors charge - I think its about $5-$6 per tube for matching - so you are paying a premium there which is included in the cost.
an expert would have to answer that, although I believe once you have the operating point covered, it should not be a problem, atleast with tubes of the same brand.
Ofcourse not. I meant seperate/individual resistors for each tube.
thats a good idea if cost is a bar.
Create one
chromal said:
or in the seperate pot/resistor arrangement, you could just change one tube and save half the cost and might still be better off than using a "matched pair". Not to mention vendors charge - I think its about $5-$6 per tube for matching - so you are paying a premium there which is included in the cost.
an expert would have to answer that, although I believe once you have the operating point covered, it should not be a problem, atleast with tubes of the same brand.
Ofcourse not. I meant seperate/individual resistors for each tube.
thats a good idea if cost is a bar.
The info out there seems minimal at best, and I have yet to come across something as simple as a grid/plate curve chart.
Create one
Re: Does tube matching make sense ?
An interesting article on minimizing that (requires 4 OPTs for stereo thou)
http://www.pearl-hifi.com/06_Lit_Archive/07_Misc_Downloads/Twinning_PushPull_OPTs.pdf
Yep, you have to replace the pair... or if you have an all-wheel drive car, all 4 (i have 2 sets of 3 good tires if anyone is interested).
Depending on how deep you are into their lifespan, you are unlikely to find somethng close enuff anyway.
dave
percy said:
An interesting article on minimizing that (requires 4 OPTs for stereo thou)
http://www.pearl-hifi.com/06_Lit_Archive/07_Misc_Downloads/Twinning_PushPull_OPTs.pdf
Yep, you have to replace the pair... or if you have an all-wheel drive car, all 4 (i have 2 sets of 3 good tires if anyone is interested).
Depending on how deep you are into their lifespan, you are unlikely to find somethng close enuff anyway.
dave
Capacitor Upgrades
I finally got around do doing some more mods on the K-12M and I am very glad I did.
As suggested by Blackie, I removed the DC blocking input capacitors. This has turned out to be the biggest improvement thus far. The best sounding capacitors are clearly no capacitors at all. Removal of these caps really reduced midrange and tweeter hiss.
I upgraded the four coupling capacitors to film and foil. This also had a big impact on further improving the sound. The sound with the film and foil caps is more dynamic. Also, all mid and high frequency is gone .. the amp is dead quiet now.
I added some polystyrene caps to the electolytic PSU capacitors. This also had a positive effect. On this kit, small changes to the PSU seem to have a big impact on sound.
Finally, I added some cheap tube dampers. These also helped the sound slightly. However, I am thinking that tube coolers (which will also provide some damping) may be a better upgrade to invest in over dampers.
Anyway, you can read more on my website.
I finally got around do doing some more mods on the K-12M and I am very glad I did.
As suggested by Blackie, I removed the DC blocking input capacitors. This has turned out to be the biggest improvement thus far. The best sounding capacitors are clearly no capacitors at all. Removal of these caps really reduced midrange and tweeter hiss.
I upgraded the four coupling capacitors to film and foil. This also had a big impact on further improving the sound. The sound with the film and foil caps is more dynamic. Also, all mid and high frequency is gone .. the amp is dead quiet now.
I added some polystyrene caps to the electolytic PSU capacitors. This also had a positive effect. On this kit, small changes to the PSU seem to have a big impact on sound.
Finally, I added some cheap tube dampers. These also helped the sound slightly. However, I am thinking that tube coolers (which will also provide some damping) may be a better upgrade to invest in over dampers.
Anyway, you can read more on my website.
volume pot in shunt mode
GG, since you already removed the DC blocking caps you may be interested in this thread on the Decware forum on running the volume pot in shunt mode with an added resistor
http://www.decware.com/cgi-bin/yabb/YaBB.cgi?board=Tubes&action=display&num=1110965291
GG, since you already removed the DC blocking caps you may be interested in this thread on the Decware forum on running the volume pot in shunt mode with an added resistor
http://www.decware.com/cgi-bin/yabb/YaBB.cgi?board=Tubes&action=display&num=1110965291
Hello everyone I am new here to the forums, I am mostly on the headfi.org...yeah I am a headphone guy.
anyway I bought this kit last week and fell in love with it.
I have another kit coming in this week so I can run these as mono's, and I already know how to do it.
the gear I am using the kit in is a Meridian G08 cdp and NORH 6.6 speakers (93db senn)...
any thing else I should be looking into as mods?
thank you and its glad to be here, I can see myself here for a long time.
anyway I bought this kit last week and fell in love with it.
I have another kit coming in this week so I can run these as mono's, and I already know how to do it.
the gear I am using the kit in is a Meridian G08 cdp and NORH 6.6 speakers (93db senn)...
any thing else I should be looking into as mods?
thank you and its glad to be here, I can see myself here for a long time.
Re: volume pot in shunt mode
Funny that you mention this. The original pot that came with my kit was starting to fail so I replaced it with an old recycled pot. Anyway, I do not like how this pot positions (normal listenting level is about 2-3 oclock) so I had already read about doing this on this passive preamp thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=2608&perpage=10&highlight=&pagenumber=2
I too have been very pleased with this kit. I was not expecting this sort of performance for $139.
As for mods, you must visit:
http://www.siteswithstyle.com/VoltSecond/K-12M_AMP/K-12M_Push_Pull.html
I have tried most of VoltSeconds mods. Those mods are very well thought out and they are well documented.
Also, you can visit my site for some additional capacitor mods.
Of the mods I have tried, I would stongly reccomend the snubbers on the filaments and the HV. Also, if your sources do not pass DC, removing the DC blocking caps made the biggest improvement of all the mods I have tried. I also found that upgrading the coupling caps to film/foil caps made a large improvement.
Good luck,
Gio.
james_b said:
Funny that you mention this. The original pot that came with my kit was starting to fail so I replaced it with an old recycled pot. Anyway, I do not like how this pot positions (normal listenting level is about 2-3 oclock) so I had already read about doing this on this passive preamp thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=2608&perpage=10&highlight=&pagenumber=2
bhd812 said:
I too have been very pleased with this kit. I was not expecting this sort of performance for $139.
As for mods, you must visit:
http://www.siteswithstyle.com/VoltSecond/K-12M_AMP/K-12M_Push_Pull.html
I have tried most of VoltSeconds mods. Those mods are very well thought out and they are well documented.
Also, you can visit my site for some additional capacitor mods.
Of the mods I have tried, I would stongly reccomend the snubbers on the filaments and the HV. Also, if your sources do not pass DC, removing the DC blocking caps made the biggest improvement of all the mods I have tried. I also found that upgrading the coupling caps to film/foil caps made a large improvement.
Good luck,
Gio.
Hi All --
This past weekend I finished my rendition of this amp, I have the k-502 from AES with 11BM8 tubes, and while I'm very happy with the reults I'm not too keen on the lack of bleeder resistors -- as a first-time DIYer, I'm suitably paranoid about it all.
Voltsecond's CCS mod is one way of rectifying that, and I was wondering if Doc B's C4S kit would work here.
Thanks in advance for your yay or nay.
John
This past weekend I finished my rendition of this amp, I have the k-502 from AES with 11BM8 tubes, and while I'm very happy with the reults I'm not too keen on the lack of bleeder resistors -- as a first-time DIYer, I'm suitably paranoid about it all.
Voltsecond's CCS mod is one way of rectifying that, and I was wondering if Doc B's C4S kit would work here.
Thanks in advance for your yay or nay.
John
Voltsecond's CCS mode for K502
John,
I have the K502 amp and recently built the CCS mod for it tailored for the K502 amp. Voltsecond was kind enough to discuss it with me and suggested some changes on a few of the components to suit the K502. The K502 uses 200ohm 3W cathode resistors where the S5 used 300ohm, plus my voltage readings across these cathode resistors was only 15V where the S5 had 25V. I corresponded with one other K502 owner who had similar voltage readings on their kit. Voltsecond recommended the following changes:
-do not use an LED
-do not use R5, R6 or R7
-for R2 & R3 use 120k 1/4W [instead of 130k]
-for R4 use 100ohm 1/4W - [instead of 110ohm]
All the other components are the same. He said that would leave enough power later for the balance mod he is working on.
This has been in my amp about two weeks now and I really like it, it does make the sound a little softer and smoother. Just use clipon heatsink clamps when you solder, especially the hex/mos fet, capacitor & diodes.
John,
I have the K502 amp and recently built the CCS mod for it tailored for the K502 amp. Voltsecond was kind enough to discuss it with me and suggested some changes on a few of the components to suit the K502. The K502 uses 200ohm 3W cathode resistors where the S5 used 300ohm, plus my voltage readings across these cathode resistors was only 15V where the S5 had 25V. I corresponded with one other K502 owner who had similar voltage readings on their kit. Voltsecond recommended the following changes:
-do not use an LED
-do not use R5, R6 or R7
-for R2 & R3 use 120k 1/4W [instead of 130k]
-for R4 use 100ohm 1/4W - [instead of 110ohm]
All the other components are the same. He said that would leave enough power later for the balance mod he is working on.
This has been in my amp about two weeks now and I really like it, it does make the sound a little softer and smoother. Just use clipon heatsink clamps when you solder, especially the hex/mos fet, capacitor & diodes.
Heh, I figured I'd weigh in here again, I've been mostly silent lately.
I replaced the 300 ohm 3w cathode resistors with 400 ohm wirewound 5w resistors. I wanted to use 500 ohm resistors, but the company I bought 'em from screwed up and included a 50 ohm resistor. I should have followed my rule of buying extra passives with each order.
Anyway, This increase in cathode resistance raised the cathode about 2v, and dropped the zero-signal current through the tubes about 1.25w each. This also caused the B+ voltage to jump up a bit, maybe 5-10v.
I hooked up a function generator and watched the L channel OT secondary across a 4ohm dummy load. Looks like the more-or-less stock amp can go up to about 45khz without any loss in gain, with a half-voltage on the output up around 75khz. (1/2 voltage == -3dB? Highpass Cutoff?) Square wave output gets a bit funky below 1000Hz, and definitely things go south below around 100Hz. I need to do some checks and make sure my function generator is healthy, because the square waves coming right out of it don't look all that square, with sloped plateaus at lower freqs, so, YMMV....
It does appear that this amp boosts upper freqs. As I sweep a tone up, the output voltage peaks tend to show some moderate (maybe up to 15-20%) magnitude gain, at least to a point.
Has anyone thought about or tried triode strapping? Does anyone know what the primary:secondary turn ratio is on the OTs?
I replaced the 300 ohm 3w cathode resistors with 400 ohm wirewound 5w resistors. I wanted to use 500 ohm resistors, but the company I bought 'em from screwed up and included a 50 ohm resistor. I should have followed my rule of buying extra passives with each order.
Anyway, This increase in cathode resistance raised the cathode about 2v, and dropped the zero-signal current through the tubes about 1.25w each. This also caused the B+ voltage to jump up a bit, maybe 5-10v.
I hooked up a function generator and watched the L channel OT secondary across a 4ohm dummy load. Looks like the more-or-less stock amp can go up to about 45khz without any loss in gain, with a half-voltage on the output up around 75khz. (1/2 voltage == -3dB? Highpass Cutoff?) Square wave output gets a bit funky below 1000Hz, and definitely things go south below around 100Hz. I need to do some checks and make sure my function generator is healthy, because the square waves coming right out of it don't look all that square, with sloped plateaus at lower freqs, so, YMMV....
It does appear that this amp boosts upper freqs. As I sweep a tone up, the output voltage peaks tend to show some moderate (maybe up to 15-20%) magnitude gain, at least to a point.
Has anyone thought about or tried triode strapping? Does anyone know what the primary:secondary turn ratio is on the OTs?
will triode strap soon
I plan on triode strapping my amp soon. I have not found any other posts where anyone has triode strapped the K12M or K502 amp. There was one post where the two OPT's were replaced with Hammond 1609's and wired in ultralinear mode with great results. I figured I would try triode strapping first as the cost of 4 resistors and some wire is much less than 2 OPTs at @ $75/pair. When I went back and looked for this link I found a new post on other mods everyone may be interested in, like replacing the 2 33pf ceramic caps with silver mica caps, so since both amps are similar here it is also.
http://diytube.com/phpBB2/viewtopic.php?t=152
http://diytube.com/phpBB2/viewtopic.php?p=3092#3092
I plan on triode strapping my amp soon. I have not found any other posts where anyone has triode strapped the K12M or K502 amp. There was one post where the two OPT's were replaced with Hammond 1609's and wired in ultralinear mode with great results. I figured I would try triode strapping first as the cost of 4 resistors and some wire is much less than 2 OPTs at @ $75/pair. When I went back and looked for this link I found a new post on other mods everyone may be interested in, like replacing the 2 33pf ceramic caps with silver mica caps, so since both amps are similar here it is also.
http://diytube.com/phpBB2/viewtopic.php?t=152
http://diytube.com/phpBB2/viewtopic.php?p=3092#3092
Just because you brought up swapping transformers... if you were change out the power transfomer, which Hammond would you pick and how would you know that that one is the right one? I'm more interested in how you'd figure out which one is the right one than actually knowing which one, if that makes any sense.
Thanks, John.
Thanks, John.
chromal said:
nullspace said:
I can't find my notes as I never built this kit - which at some point I was going to. But I believe the OPT primary Z is 10k and yes I was going to use the Hammond 1609. I cant remember the power transformer but measure the ac voltage across the secondary and the total current drain of the amp, that should give you an idea of the rating you need. Same for the heater winding/transformer.
Hi All --
I picked up four of the FREDs that Blackie mentioned back when this thread was getting started, and I plan on doing something with them this weekend.
However, I was wondering if someone has a better idea than mine on building the bridge. I was planning on juggling the four FREDs so that the anodes and cathodes lined up, then running a lead from each of the four points back to the PCB. Will this work, or is there a better way to build a rectifier from TO-220AC diodes?
Thanks for the help,
John
I picked up four of the FREDs that Blackie mentioned back when this thread was getting started, and I plan on doing something with them this weekend.
However, I was wondering if someone has a better idea than mine on building the bridge. I was planning on juggling the four FREDs so that the anodes and cathodes lined up, then running a lead from each of the four points back to the PCB. Will this work, or is there a better way to build a rectifier from TO-220AC diodes?
Thanks for the help,
John
killerfishes said:
The unused pads are most likely meant for bypass caps for the common cathode resistors (R8 and R17) for the push-pull pairs.
Adding a pair of 50uf 25V caps will improve the bass response, although some people might cringe with the mere thought of adding an electrolytic cap into the signal path.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.