Scott LK-72 Rehab
- Thread starter Dave451
- Start date
Dave451, is the schematic clear enough for you? It looks blurry when l open my One Drive link... I'll try to link the original back in 2009 or 2010.
I'll try to look tomorrow on my computer if the jpeg is better looking, didn't find the exact post or PM, time to go to bed!
I'll try to look tomorrow on my computer if the jpeg is better looking, didn't find the exact post or PM, time to go to bed!
Last edited:
Patrice, the schematic is fine; very legible and thanks for forwarding. When you first call it up and try to increase size, it is blurry before it settles on the expanded image and the clarity returns.
Dave G: thanks for your feedback on using the no. 1 7591 pin as a tie point. I'll proceed that way when I have the resistors in hand.
Dave G: thanks for your feedback on using the no. 1 7591 pin as a tie point. I'll proceed that way when I have the resistors in hand.
Pretty much completed the electrical rehab work for the LK-72 last 2-3 days after the parts shipment arrived. Completed everything except for a physical cleanup of the chassis, front panel and knobs, including:
1. Cleaned the tube socket pins, pots, rotary and slide switches with DeOxit D5 for the switches and FaderLube for the pots. Also cleaned the RCA connectors and the speaker terminal strips
2. Carefully removed the can caps except for the 4 x 75 uF bias/filament supply filter can. Epoxied three new F&T 2X 32 uF/500 VDC caps to the phenolic wafers that had held the Astron can caps. Wired in the new caps per prior post. One note: you have to be careful when working on this unit to be sure the ground paths are preserved. The ground system is basically a "daisy chain" bus that runs throughout the chassis and is connected to a couple of chassis points. None of the can caps had the can physically soldered to the chassis; all were interconnected with wires forming the ground bus. I was thoughtful about how I did this to avoid introducing ground loops and noise. (photos)
3. Replaced the 4 x 75 uF/75 VDC bias can with four axial IC 100 uF/100V caps. I removed the one point tie strip by the power transformer and installed a new 3 point tie strip with 2 ground points, along with a single on one a transformer screw to accommodate the bias/filament DC supply without any flying leads. All the original 18 ohm resistors from the R-C string tested good, so I reused three of them, along with the 3W 30 ohm metal film resistor I used instead of the fourth 18 ohm 1W right after the rectifier bridge.
4. I installed four red test points near the cathodes of the 7591A output tubes to enable checking the bias above chassis and connected with short wire jumpers. Also mounted a black common negative test point nearby (photos). Because I was going to install 10 ohm cathode resistors, I had removed the 7591 cathode jumpers and, with respect to the ground continuity observation above, installed a ground point under a transformer screw with a star washer, to which I grounded the negative test point, one of the 10 ohm cathode resistors, and the ground point for the 68 ohm resistors in the feedback circuit.
5. Installed 10 ohm 1/2 watt precision metal film resistors on each cathode and grounded each to the nearest ground point.
6. Re-routed the 7591 screen supply to pin 1 of each 7591, removed the inter-tube screen jumpers and left in place the pin 1 to pin 8 screen jumpers on each tube. Installed 100 ohm metal film screen protection resistors between pin 1 and pin 8 on each tube socket. You can see these on the photos.
7. Drilled and mounted two 10K linear taper bias pots between the 12AX7's, basically the only place I could find to mount good-quality pots for top-of-chassis adjustment of bias level. Wired the bias level controls per the schematic provided by Patrice in prior post in this thread with 100 uf 100V electrolytics wired from the new bias control wipers to ground (positive grounded, of course). Cut the bias jumper between the DC balance pots and wired the wipers of the new bias level pots to the wipers of the respective DC balance pots, so now I have a way to adjust both bias level and balance for each side to set a bias level for each tube using the new test points.
8. I replaced all the Ceracap 0.1 uf 400V couplers with Solen Fast Cap 0.1 uF 630V units.
After triple-checking the wiring and new components, I powered up the bias/filament DC supply with only the 12AX7's in place and all checked out well. With a little experimenting, I found that using a 22K resistor in place of the 10K in Patrice's schematic gave me a spread of about 3.5 VDC and I measured from about -18.5VDC to around -15V swing at the tube pins when adjusting the level pot through its full travel.
I put in the rest of the tubes including the 5AR4 rectifier, hooked up speakers and brought power up fairly slowly on the Variac to ease in the new can caps. All was well with B+, plate and screen voltages on the tubes at full line voltage. I have not yet installed a CL-80, as I don't have any on hand, but that will be done before I finish up.
I checked the bias on the tubes and all were fully responsive to the new bias adjustment circuit. Unfortunately, it looks like my original Sylvania 7591A's are getting on--there appears to be one strong tube and one weak in each pair. I was able to balance them fairly closely, but one tube needed around -16.5 volts while the other needed -18.5 volts. At the maximum negative voltage setting on the level pot, the tubes were still drawing around 40-45 mA each and I'm looking for 33 to 35 mA, no more than 38 mA per tube.
I re-balanced the DC balance pots and kept the bias level pots at max negative voltage and tried a new set of Tung-Sol tubes and these were drawing 54 to 56 mA each (!), so I shut them down quickly and put the old ones back in and balanced them for a sound check. The sound was perfectly good with the old tubes and I'm happy to report no buzz or hum, including in the phono mode--just a small amount of tube rush from the phono stage (no ground loops!?).
So, we're in business, but I have to get more negative bias voltage available at the tube pins. Changing out the 'tail' resistor between the bias level pots and ground doesn't do much to increase the overall bias level, it mainly just changes the span of control. After studying, diagramming, and doing calculations on the bias divider network until my head hurt, I'm thinking the only way to raise the overall bias level to the tubes is to change the 270K resistors in the grid circuit to something around 180K and that should do it (get me maybe -20 or -20.5 VDC at max negative). With the bias level pots in the max negative voltage setting, I'm getting the full -45-47VDC to the wipers of the DC balance pots, but still not enough negative bias voltage at the grids to cool down the tubes, apparently.
QUESTION: will changing the resistors in this way result in any stability issues with the 7591 tubes or with the drive level from the driver/PI 6GH8A's (or any other issues)?
So, the pictures show the 'after rehab' state with only a good physical clean up and resolving the bias issue to be complete.
Dave
Pictures: 1) overall underneath after the work, 2) bias/filament supply re-wired with new bridge rectifier, 3) detail of new bias level pots & wiring, 4) Wiring around 7591 tube sockets, including bias test points, 5) Common negative test point and new ground point, 6) Overall view topside, 7) close up of test points, 8) All test points and, 9) new bias level pots, adjustable from above or below deck.
1. Cleaned the tube socket pins, pots, rotary and slide switches with DeOxit D5 for the switches and FaderLube for the pots. Also cleaned the RCA connectors and the speaker terminal strips
2. Carefully removed the can caps except for the 4 x 75 uF bias/filament supply filter can. Epoxied three new F&T 2X 32 uF/500 VDC caps to the phenolic wafers that had held the Astron can caps. Wired in the new caps per prior post. One note: you have to be careful when working on this unit to be sure the ground paths are preserved. The ground system is basically a "daisy chain" bus that runs throughout the chassis and is connected to a couple of chassis points. None of the can caps had the can physically soldered to the chassis; all were interconnected with wires forming the ground bus. I was thoughtful about how I did this to avoid introducing ground loops and noise. (photos)
3. Replaced the 4 x 75 uF/75 VDC bias can with four axial IC 100 uF/100V caps. I removed the one point tie strip by the power transformer and installed a new 3 point tie strip with 2 ground points, along with a single on one a transformer screw to accommodate the bias/filament DC supply without any flying leads. All the original 18 ohm resistors from the R-C string tested good, so I reused three of them, along with the 3W 30 ohm metal film resistor I used instead of the fourth 18 ohm 1W right after the rectifier bridge.
4. I installed four red test points near the cathodes of the 7591A output tubes to enable checking the bias above chassis and connected with short wire jumpers. Also mounted a black common negative test point nearby (photos). Because I was going to install 10 ohm cathode resistors, I had removed the 7591 cathode jumpers and, with respect to the ground continuity observation above, installed a ground point under a transformer screw with a star washer, to which I grounded the negative test point, one of the 10 ohm cathode resistors, and the ground point for the 68 ohm resistors in the feedback circuit.
5. Installed 10 ohm 1/2 watt precision metal film resistors on each cathode and grounded each to the nearest ground point.
6. Re-routed the 7591 screen supply to pin 1 of each 7591, removed the inter-tube screen jumpers and left in place the pin 1 to pin 8 screen jumpers on each tube. Installed 100 ohm metal film screen protection resistors between pin 1 and pin 8 on each tube socket. You can see these on the photos.
7. Drilled and mounted two 10K linear taper bias pots between the 12AX7's, basically the only place I could find to mount good-quality pots for top-of-chassis adjustment of bias level. Wired the bias level controls per the schematic provided by Patrice in prior post in this thread with 100 uf 100V electrolytics wired from the new bias control wipers to ground (positive grounded, of course). Cut the bias jumper between the DC balance pots and wired the wipers of the new bias level pots to the wipers of the respective DC balance pots, so now I have a way to adjust both bias level and balance for each side to set a bias level for each tube using the new test points.
8. I replaced all the Ceracap 0.1 uf 400V couplers with Solen Fast Cap 0.1 uF 630V units.
After triple-checking the wiring and new components, I powered up the bias/filament DC supply with only the 12AX7's in place and all checked out well. With a little experimenting, I found that using a 22K resistor in place of the 10K in Patrice's schematic gave me a spread of about 3.5 VDC and I measured from about -18.5VDC to around -15V swing at the tube pins when adjusting the level pot through its full travel.
I put in the rest of the tubes including the 5AR4 rectifier, hooked up speakers and brought power up fairly slowly on the Variac to ease in the new can caps. All was well with B+, plate and screen voltages on the tubes at full line voltage. I have not yet installed a CL-80, as I don't have any on hand, but that will be done before I finish up.
I checked the bias on the tubes and all were fully responsive to the new bias adjustment circuit. Unfortunately, it looks like my original Sylvania 7591A's are getting on--there appears to be one strong tube and one weak in each pair. I was able to balance them fairly closely, but one tube needed around -16.5 volts while the other needed -18.5 volts. At the maximum negative voltage setting on the level pot, the tubes were still drawing around 40-45 mA each and I'm looking for 33 to 35 mA, no more than 38 mA per tube.
I re-balanced the DC balance pots and kept the bias level pots at max negative voltage and tried a new set of Tung-Sol tubes and these were drawing 54 to 56 mA each (!), so I shut them down quickly and put the old ones back in and balanced them for a sound check. The sound was perfectly good with the old tubes and I'm happy to report no buzz or hum, including in the phono mode--just a small amount of tube rush from the phono stage (no ground loops!?).
So, we're in business, but I have to get more negative bias voltage available at the tube pins. Changing out the 'tail' resistor between the bias level pots and ground doesn't do much to increase the overall bias level, it mainly just changes the span of control. After studying, diagramming, and doing calculations on the bias divider network until my head hurt, I'm thinking the only way to raise the overall bias level to the tubes is to change the 270K resistors in the grid circuit to something around 180K and that should do it (get me maybe -20 or -20.5 VDC at max negative). With the bias level pots in the max negative voltage setting, I'm getting the full -45-47VDC to the wipers of the DC balance pots, but still not enough negative bias voltage at the grids to cool down the tubes, apparently.
QUESTION: will changing the resistors in this way result in any stability issues with the 7591 tubes or with the drive level from the driver/PI 6GH8A's (or any other issues)?
So, the pictures show the 'after rehab' state with only a good physical clean up and resolving the bias issue to be complete.
Dave
Pictures: 1) overall underneath after the work, 2) bias/filament supply re-wired with new bridge rectifier, 3) detail of new bias level pots & wiring, 4) Wiring around 7591 tube sockets, including bias test points, 5) Common negative test point and new ground point, 6) Overall view topside, 7) close up of test points, 8) All test points and, 9) new bias level pots, adjustable from above or below deck.
Attachments
this schematic right?
https://onedrive.live.com/?authkey=...C0&id=51375E3DFB49FC0!6352&parId=root&o=OneUp
So the problem is you need more negative voltage right? Changing the grid resistor really isn't a good way to do that. What you want is to reduce the voltage at the pots. Changing the 10K resistor to ground off the tail of the pots will alter the range of what they can deliver. The higher the value, the less range the pots will have. Its not going to give you appreciably more negative voltage though. Try moving the feed to the bias pots to the other side of R207. That will increase the amount of negative voltage fed to the pots. If thats not enough, try it on the other side of R208. Once you get enough negative voltage you may want to play with the value of that 10K resistor to keep the adjustment range of the pots useful. Too low a resistor value and it will become overly touchy.
the Sylvanias don't sound like they're in awful shape honestly. The new production tubes also do seem to run pretty hot vs the old ones so you may find that you want more total range just to allow both types to be used going forward.
https://onedrive.live.com/?authkey=...C0&id=51375E3DFB49FC0!6352&parId=root&o=OneUp
So the problem is you need more negative voltage right? Changing the grid resistor really isn't a good way to do that. What you want is to reduce the voltage at the pots. Changing the 10K resistor to ground off the tail of the pots will alter the range of what they can deliver. The higher the value, the less range the pots will have. Its not going to give you appreciably more negative voltage though. Try moving the feed to the bias pots to the other side of R207. That will increase the amount of negative voltage fed to the pots. If thats not enough, try it on the other side of R208. Once you get enough negative voltage you may want to play with the value of that 10K resistor to keep the adjustment range of the pots useful. Too low a resistor value and it will become overly touchy.
the Sylvanias don't sound like they're in awful shape honestly. The new production tubes also do seem to run pretty hot vs the old ones so you may find that you want more total range just to allow both types to be used going forward.
Dave -- Super work!! -- But a couple of points if I may:
1. The 330K resistors originally connected from pin 6 to ground (grounded by way of pin 5 which used to be directly grounded) still need to be grounded, but now they are not -- they are connected to the cathode that now operates slightly above ground. It's not a huge matter, but a point of detail none the less. That's the part of the Scott design that makes it less than easy to install DC balance controls, because once the cathodes are no longer grounded, there's no convenient ground terminal to use for those resistors. That was why I finally resorted to using the 14 ga ground wire snaking through the output tube area to provide all the necessary ground points as I showed in the thread on the 240 volt damaged power transformer unit I posted on recently. Again, not a big deal, but something to be aware of......
2. You cannot judge the condition of a power tube based on how much bias voltage it does or does not require to produce a given current draw -- not unless it's sooo bad that it takes a notably less amount -- like say -8 volts or so (exaggerated to make the point). But with just a 2 volt bias spread that is still within the realm of normalcy for these tubes, that could be a weak cathode, or it could be just a tube with wider manufacturing tolerances in its construction causing the different bias voltage requirement. This is why I've always said that the only way you'll truly know anything about your power tubes is to do a power output test on them. That doesn't lie. Anything else is just an indication of characteristic -- not condition.
Dave
1. The 330K resistors originally connected from pin 6 to ground (grounded by way of pin 5 which used to be directly grounded) still need to be grounded, but now they are not -- they are connected to the cathode that now operates slightly above ground. It's not a huge matter, but a point of detail none the less. That's the part of the Scott design that makes it less than easy to install DC balance controls, because once the cathodes are no longer grounded, there's no convenient ground terminal to use for those resistors. That was why I finally resorted to using the 14 ga ground wire snaking through the output tube area to provide all the necessary ground points as I showed in the thread on the 240 volt damaged power transformer unit I posted on recently. Again, not a big deal, but something to be aware of......
2. You cannot judge the condition of a power tube based on how much bias voltage it does or does not require to produce a given current draw -- not unless it's sooo bad that it takes a notably less amount -- like say -8 volts or so (exaggerated to make the point). But with just a 2 volt bias spread that is still within the realm of normalcy for these tubes, that could be a weak cathode, or it could be just a tube with wider manufacturing tolerances in its construction causing the different bias voltage requirement. This is why I've always said that the only way you'll truly know anything about your power tubes is to do a power output test on them. That doesn't lie. Anything else is just an indication of characteristic -- not condition.
Dave
kvflyer
Super Member
One comment that I may make which is in no way a criticism. What I have done when replacing the electrolytic capacitors is to simply drill out the phonelic wafer rivets and replace it with a clamp. I have found the clamp holes usually line up with the holes in the chassis. Since the replacement capacitors don't look original anyway, that eliminates the necessity of gluing them down. It also allows future repairs easily (which hopefully we won't have to do). And clamps are inexpensive
Surely we all have seen clamped capacitors in the Japanese amplifiers, Pioneer, Sansui, Kenwood etc. yes, I would prefer to see original-looking aluminum cans. But the clamped capacitors look very professional. And easy to do. There are a few manufacturers out there which make the 50-50 or 32-32 Mfd capacitors with excellent specs.
However... I if at all possible, I uncrimp the original electrolytic capacitor, pull out the old "guts" and restuff them. I did a thread on that process a few years back if you are interested.
Again, please, not a criticism. Just another approach if I may.
Surely we all have seen clamped capacitors in the Japanese amplifiers, Pioneer, Sansui, Kenwood etc. yes, I would prefer to see original-looking aluminum cans. But the clamped capacitors look very professional. And easy to do. There are a few manufacturers out there which make the 50-50 or 32-32 Mfd capacitors with excellent specs.
However... I if at all possible, I uncrimp the original electrolytic capacitor, pull out the old "guts" and restuff them. I did a thread on that process a few years back if you are interested.
Again, please, not a criticism. Just another approach if I may.
Guys, thanks for all the excellent and timely feedback--I really appreciate it!
Patrice, thanks again for the set up for the bias network. As you can see, it's very do-able and works quite well. I just need to get the right level of bias voltage for the tubes using the existing divider network.
Gadget, yep--that schematic is the one I used. Great idea for tapping the RC string upstream a bit to get more negative voltage! With the filtering from all those 100 uF caps, should be no worry on ripple. I'll try that as my next move. As you could see, I am nervous about doing anything around the grid resistance divider network, as it seems carefully designed and I don't want to generate collateral damage messing around with it. This seems like a great alternative.
Dave G--great catch on the 330K grid resistor grounding. That went completely over my head during the process. I'll get them properly grounded. I like your idea of the ground bus, so that may be in the LK-72's future, then I can relocate the grounds of the 10 ohm cathode resistors as well a little closer to the tubes.
As for the tube quality, I guess I was too hasty! I'd absolutely love to keep those Sylvania's going in this amp. Looking at the cathode current now, it makes me wonder how much current they were drawing with the stock design and running about -17 to -17.4 volts at the grids. Must have been pretty heavy.
Kvflyer, no way do I take this as criticism! In fact, I have used clamps with JJ's and F&T caps before. I usually use clamps when the replacements have no mounting tabs (like these) and there is no phenolic wafer. I don't think they look bad at all (see my Sherwood S8000 IV thread for a late example of a clamp-mounted JJ), but it's just a visual preference for the glued down method--looks a little cleaner to me if I can do it. As for stuffing caps, one day I'm going to bite the bullet and do it, but haven't so far. I have looked at your thread and it is excellent--when I do it, I'll do it that way.
Going to be out of town for a couple weeks, but when I return I'll work on these things and report back.
Again, thanks very much.
Dave
Patrice, thanks again for the set up for the bias network. As you can see, it's very do-able and works quite well. I just need to get the right level of bias voltage for the tubes using the existing divider network.
Gadget, yep--that schematic is the one I used. Great idea for tapping the RC string upstream a bit to get more negative voltage! With the filtering from all those 100 uF caps, should be no worry on ripple. I'll try that as my next move. As you could see, I am nervous about doing anything around the grid resistance divider network, as it seems carefully designed and I don't want to generate collateral damage messing around with it. This seems like a great alternative.
Dave G--great catch on the 330K grid resistor grounding. That went completely over my head during the process. I'll get them properly grounded. I like your idea of the ground bus, so that may be in the LK-72's future, then I can relocate the grounds of the 10 ohm cathode resistors as well a little closer to the tubes.
As for the tube quality, I guess I was too hasty! I'd absolutely love to keep those Sylvania's going in this amp. Looking at the cathode current now, it makes me wonder how much current they were drawing with the stock design and running about -17 to -17.4 volts at the grids. Must have been pretty heavy.
Kvflyer, no way do I take this as criticism! In fact, I have used clamps with JJ's and F&T caps before. I usually use clamps when the replacements have no mounting tabs (like these) and there is no phenolic wafer. I don't think they look bad at all (see my Sherwood S8000 IV thread for a late example of a clamp-mounted JJ), but it's just a visual preference for the glued down method--looks a little cleaner to me if I can do it. As for stuffing caps, one day I'm going to bite the bullet and do it, but haven't so far. I have looked at your thread and it is excellent--when I do it, I'll do it that way.
Going to be out of town for a couple weeks, but when I return I'll work on these things and report back.
Again, thanks very much.
Dave
Any ripple goes in common mode anyway. The bias supply can be pretty noisy and you'd never hear it at the output. Same signal into both grids will cancel. I've put the scope on some other things that were fixed bias, you might be surprised how much ripple some of them have with stock design and stock value parts.
triode17
Super Member
I second the motion.
triode17
Super Member
I second the motion.
triode17
Super Member
Dave 451 - You said this, "Before, I have installed a couple of tie strips and run a bus for the screen voltage supply, taking the 100 ohm resistors off the bus to the screen pins."
This could have led to a disaster. The screen grid (or any grid ) resistors should be mounted as close to the tube itself as possible. This is to prevent the oscillations that the 100 Ohm resistors are there to stop in the first place.
This could have led to a disaster. The screen grid (or any grid ) resistors should be mounted as close to the tube itself as possible. This is to prevent the oscillations that the 100 Ohm resistors are there to stop in the first place.
triode17
Super Member
Dave 451 - You said, "So, we're in business, but I have to get more negative bias voltage available at the tube pins. Changing out the 'tail' resistor between the bias level pots and ground doesn't do much to increase the overall bias level, it mainly just changes the span of control. After studying, diagramming, and doing calculations on the bias divider network until my head hurt, I'm thinking the only way to raise the overall bias level to the tubes is to change the 270K resistors in the grid circuit to something around 180K and that should do it (get me maybe -20 or -20.5 VDC at max negative). With the bias level pots in the max negative voltage setting, I'm getting the full -45-47VDC to the wipers of the DC balance pots, but still not enough negative bias voltage at the grids to cool down the tubes, apparently."
When I rebuilt my LK72 (A), I replaced the selenium but reset the series string to produce 48 v. across V1-V4 heaters. This way, they run at 12.0v, not the previous 11.25, which to me is too low. Then, you should have the extra negative bias voltage you need.
When I rebuilt my LK72 (A), I replaced the selenium but reset the series string to produce 48 v. across V1-V4 heaters. This way, they run at 12.0v, not the previous 11.25, which to me is too low. Then, you should have the extra negative bias voltage you need.
PatR
New Member
Hope I'm not hijacking, but since this is a Scott LK-72 thread I thought it would be okay to just ask here as opposed to starting a whole new thread for a simple question. I have an LK-72a that I am about to start rebuilding. Lots of good tips here in this thread for sure.. My question. After glancing over the schematic (and I know from experience with a 299c/233 schematic that Scott Schematics are not reliable) I see C25/125 are listed as .25 lytics (polarity noted on schematic). Of course finding a polarized .25 cap looks to be impossible. There is nothing even close as mouser or AES. Can anyone here with experience rebuilding one of these amps make a suggestion for what to put in that position? Value/Type? Thanks!
As I recall, that is actually either a film cap or a ceramic. I believe I used a .27 film cap that was rated 250V AC. You should probably start another thread since you may have a bunch more questions There is some real nice info here for rebuilding that amp. I think a lot of people use a .22uf there also. I don't think the value is extremely critical.
Last edited:
Pat -- The schematic you're looking at is deceiving because the + and - marks for those caps are left over from the previous version that used a 4 uF electrolytic cap in that position. On the schematic version you're using, the value was changed to .25 uF, but somebody missed the memo to also remove the polarity indications when that change in value was made. Any good quality mylar value will work just fine here.
Dave
Dave