Luckiestmanalive Posted July 5, 2022 Author Posted July 5, 2022 (edited) On 03/07/2022 at 9:18 PM, Luckiestmanalive said: On the three resistors I said I'd have to check close up: 1. The extra resistor on my amp connected to the left power cap is a 33Kohm resistor (orange orange orange). 2. There is a difference between a resistor connected to the left power cap - on your amp it is a 150Kohm resistor (Brown Green Yellow - R307?) and on mine it is 27Kohms (Red Violet Orange). 3. The other one is a 240Kohm resistor (Red Yellow Yellow - could be R303) on both our amps, which matches the value of the circuit diagram supplied with the conversion instructions but its value is 330Kohms on the stock MQ3600 and A3600 circuit diagrams. Oh, and there is a 10Kohm resistor (Brown Black Orange) in the preamp socket that I can't compare with yours. I've had the amp upside down on the bench to look more closely at what is going on in the connections to the wire from the bias controls. My assessment of the second difference (see above) wasn't correct. A closer look revealed a well-hidden 150Kohm resistor connected, which matches yours. The 27Kohm resistor is another extra! I found it useful to examine the doodle of this part of the circuit on the schematic I received with the amp. I checked my amp and it is mostly as drawn. However, I've added the extra resistors (27K and 33K) and the preamp socket 10Kohm resistor to make it complete (see below). Tracing the circuit from the power cap on the right there are two paths to the lead to the bias controls... Edited July 5, 2022 by Luckiestmanalive
Luckiestmanalive Posted July 5, 2022 Author Posted July 5, 2022 (edited) Ah, OK, so the combination of the resistance in the two paths is 33Kohms on one and 37Kohms (27K+10K) on the other, making for roughly 17KOhms? I guess measurement of resistance is complicated by the capacitor bank (the reading is about 8-9Kohms)? I figure I have three choices: 1. Preserve the wiring to the preamp socket by removing the extra resistors and replacing the 10Kohm resistor in the preamp socket with a 20Kohm one. 2. Remove the connection to the preamp socket and replace the 33Kohm resistor with a 20Kohm one. 3. Replace the 27Kohm resistor with a 10Kohm one and remove the 33Kohm resistor. Option 1 would make the circuit closer to the original schematic plus conversion instructions so I'm leaning toward that one. Edited July 5, 2022 by Luckiestmanalive
Luckiestmanalive Posted July 5, 2022 Author Posted July 5, 2022 (edited) Um, I've realised I left out of the diagram above a bare strip of wire that connects terminals 1 and 8 on the outside of the preamp socket. Should that be there? Yes, it should - I can see the reference to 1 and 8 on the circuit diagram near the AC power inlet. Edited July 7, 2022 by Luckiestmanalive Answered my own question
Luckiestmanalive Posted July 5, 2022 Author Posted July 5, 2022 Also, apologies, Kirk, for the lack of care and patience I've been showing. I've re-read this thread and realise I missed some important nuggets of information you have provided. I'm very grateful you have stuck with me as far as you have so far. 1
Luckiestmanalive Posted July 6, 2022 Author Posted July 6, 2022 (edited) As far as I can make out from careful examination of the picture of xlr8r's amp, it has only two connections to the bias circuit - 820ohm (the original resistor that is supposed to be changed up to a 8.2kohm resistor according to the KT88 conversion instructions) in parallel with 10kohm (value of resistor between 3 and 7 of the preamp socket) +8.2kohm in series. I've used an online resistance calculator (https://www.digikey.com.au/en/resources/conversion-calculators/conversion-calculator-parallel-and-series-resistor) to determine that these two combinations create 9kohms resistance just before the bias controls (which can add up to 30Kohms resistance). My amp has three paths to the bias lead connected to the negative terminal of the 100uf capacitor through three sets of resistors (27kohm+10kohm and 33kohm in parallel and then 8.2kohm in parallel). The 33kohm and 27kohm resistors are not in the original circuit nor in the conversion instructions. These combinations together create only 5.6Kohms resistance. If my calculations are correct, then just taking out the 33kohm resistor would increase the resistance to 6.7kohms. However, the 'official' Luxman circuit (including conversion to KT88) would be different again. The two sets of resistors would be 8.2kohm (replaces 820ohm resistor in conversion instructions) and 20kohm (replace the 10kohm resistor in the bias circuit with a 20kohm resistor). The combined resistance would be 5.8kohms. Edited July 7, 2022 by Luckiestmanalive
Luckiestmanalive Posted July 7, 2022 Author Posted July 7, 2022 (edited) Bugger - my dumb logic is wrong as it suggests that, before conversion, the two paths to the bias controls would contain an 820ohm resistor on one and a 10kohm resistor on the other (from positive terminal of 200V capacitor) - ie, the form xlr8r's amp was in before he added the 8.2kohm resistor - and this would make for a combined resistance of 4.5kohms, which was too low to increase the bias sufficiently for KT88s. Something is wrong because I cannot reduce bias enough with resistance of 5.6kohms. I'm going to have to go back to the drawing board (which may just result in me giving up and taking it to a qualified tech). Edited July 7, 2022 by Luckiestmanalive 1
Luckiestmanalive Posted July 14, 2022 Author Posted July 14, 2022 So I've sat on the bias circuit voltage problem for a while and have ordered a few 2W resistors of various values to cover my bases. The parallel paths between the positive and negative terminals of the capacitor create 17.442Kohm resistance (27k+10K on one path and 33k on the other). If I pull the 33kohm resistor and replace the 27kohm resistor with a 10kohm resistor then there will be 20kohm resistance (10k+10k). That would make the circuit as per the Luxman conversion instructions. However, I also have 8.2kohm (creating 18.2kohm resistance) and 6.8kohm (creating 16.8kohm resistance) resistors in case I need to go lower and 12k and 15k in case I need to go higher (at 10c a pop, why not?).
Luckiestmanalive Posted July 15, 2022 Author Posted July 15, 2022 (edited) OK. So my resistors arrived and I removed the 33kohm resistor that is connected to the capacitor in parallel with the path to the preamp socket and back. Then I removed the 27kohm resistor on that path and replaced it with a 10kohm and then an 8.2kohm resistor to test the difference in bias range. Here is the current the circuit provides the output tubes at the bottom end of the bias control range: A. 33kohm and 27k+10kohm in parallel (17.442kohm) = 62-63ma B. 8.2k+10kohm in series (18.2kohm) = 59-60ma C. 10K+10kohm in series (20kohm) = 54-55ma I have a 22kohm resistor and I imagine that would reduce it to ~48-49ma. There is a diagram of a MQ3600 circuit where the owner used a 22kohm resistor: https://bornschein.one/luxman_mq3600.html I'm tossing up whether to do that or leave it as it is for now. This has been a great exercise for me. I hope it provides other A3600/MQ3600 owners some value in the future. I have confirmed the Luxman instructions for converting the A3600/MQ3600 from 8045G output tubes to the KT88/6550 family of tubes. Thanks, Kirk, for providing the link on the first page. Here it is again: https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Luxman-LX-33-Service-Manual.pdf Now I'm sure the diagram that shows the 20kohm resistor connecting the positive and negative terminals (C305 to C307) on the 100uf 200V capacitor is needed only if there isn't already a connection between those same terminals routed through the preamp socket, which contains a 10kohm resistor. You can either replace that resistor with a 20kohm one or replace a wire connecting it with another 10kohm resistor (which is the path I chose). Edited July 15, 2022 by Luckiestmanalive 1
Luckiestmanalive Posted July 16, 2022 Author Posted July 16, 2022 (edited) Here are my Luxman A3600 circuit diagrams and PCB diagram with amendments shown for the KT88 (ultralinear) conversion. Edited July 16, 2022 by Luckiestmanalive 1
Luckiestmanalive Posted July 17, 2022 Author Posted July 17, 2022 (edited) On the PCB diagram, which I redrew because the one in the instructions is difficult to read, the values in red are the new values - the original values are in brackets. Note that the changes are for Ultralinear connection, not Triode connection (though some changes are common). Edited July 18, 2022 by Luckiestmanalive
Luckiestmanalive Posted July 18, 2022 Author Posted July 18, 2022 So I am very pleased with the results of my efforts. This little amp makes beautiful music without stressing the output tubes quite so much. Now I just have to research and decide what set of output tubes to buy while I'm saving - 6550s, KT88s, KT90s, or KT120s... 2
Luckiestmanalive Posted August 1, 2022 Author Posted August 1, 2022 I've updated the PCB diagram... 1
Luckiestmanalive Posted August 18, 2022 Author Posted August 18, 2022 (edited) OK, so I have been googling images for "Luxman A3600" and picking pictures of the insides to see what work other people have done to their amps. I found a series of blogposts from a guy in Japan who restored his A3600. One post includes a list of instructions to convert the A3600 from output tubes from 8045Gs to KT88s and the driver stage from 6240Gs to 6AQ8s. It has some different steps than the set above. What is particularly interesting is this list includes the 27K and 33K ohm resistors that my amp had and I removed! Google translated the kanji from this webpage: https://limited-junkroom.hatenablog.com/entry/2021/08/21/190000 I've tidied them up and I think this is a reasonable English translation: Change driver stage from 6240G to 6AQ8 1. Cathode resistance of driver tube changed from 8200Ω 2W to 12kΩ 2W 2. Change constant due to power tube change: a. Add 22kΩ 1/4W between VR of input circuit and first stage input capacitor b. Remove the cathode capacitors C103 & C104 to reduce the first stage gain c. Resistance of NF circuit (C113) - R135 & R136 from 22Ω 3W→11Ω 3W d. Power supply circuit - R305 from 820Ω →8.2kΩ and R301 from 1kΩ →2.2kΩ e. Bias power supply circuit - add 33kΩ 1W to C307 and add 27kΩ 1/2W between power connector terminal 7 and bias power supply f. Remove decoupling resistor 240kΩ in power supply circuit Change power tube from 8045G to KT88 1. Wiring addition of SG circuit Wiring between OPT/SG tap terminal and No.4 terminal of GT socket 2. In the case of triode connections, connect 4 points with 100Ω 1/2W between the 3rd and 4th terminals of the GT socket. 3. Change the NF resistance of R133 & R134 from 2.7kΩ 1/2W → 2.2kΩ 1/2W 4. The non-signal current of the KT88 is adjusted to 40mA, but initially it is 30mA or less and aged for 30 minutes or more, and then finally, the current is adjusted. * If you neglect this, even the KT88 may become red hot after a certain period of time. In terms of performance, UL connection, 1kHz 50W 0.3% 1kHz 30W 0.27% 10kHz 50W 0.6% 10kHz 30W 0.4% Fixed * In the case of my A3600, the driver tube was left as 6240G, and the DC balance resistance was changed (68k→39kΩ), which was not specified in the manual. It seems that it was left to the discretion of the site on a case-by-case basis. Edited August 18, 2022 by Luckiestmanalive
Luckiestmanalive Posted August 18, 2022 Author Posted August 18, 2022 (edited) For comparison, here is the original set of instructions, which my A3600 now conforms with (UL): For UL (Ultra Linear) Connection with KT88 (1) Change the value of R133 and R134 from 2,700-ohm 1/4W to 2,200-ohm 1/4W due to the change in NF amount (2) Change the value of C113 and C114 styrol condensers from 560pF to 330pF for stability (3) Change the value of R135 and R136 from 22-ohm 3W to 11-ohm 3W (a resistor of 22-ohm 2W can be fixed in parallel with the resistor of 22-ohm 3W) for stability (4) Add a resistor of 22K-ohm 1/2W between the center terminal of the level control volume and the circuit board for stability (5) Change the value of R305 from 820-ohm to 8.2K-ohm 1/2W for change of bias voltage (6) Change a resistor in the bias circuit from 10K-ohm 2W to 20K-ohm 2W for change of bias voltage (7) Add wires to connect in between the output SG taps and the output tube SG (4 wires) (8) Adjustment of non-signal current: Adjust to make cathode current 50mA per tube For Triode Connection with KT88 (1) Change the value of C113 and C114 styrol condensers from 560pF to 330pF (2) Change the value of R135 and R136 from 22-ohm 3W to 11-ohm 3W (or a resistor of 22-ohm 2W can be fixed in parallel) (3) Add to fix resistors, 22K-ohm 1/2W in between the center terminal of the level control volume and PCB (4) Change the value of R305 from 820-ohm to 8.2K-ohm 1/2W (5) Change a resistor in the bias circuit from 10K-ohm 2W to 20K-ohm 2W for change of bias voltage (6) Wiring of Output Tube SG Circuit: Connect with 100-ohm 1/2W resistors in between Pin 3 and Pin 4 of the sockets for triode connection of the output tubes (7) Change the value of R115 and R116 from 33K-ohm 2W to 30K-ohm 2W for improvement of distortion (8) Adjustment of non-signal current: Adjust to make cathode current 50mA per tube Edited August 18, 2022 by Luckiestmanalive
Luckiestmanalive Posted September 4, 2022 Author Posted September 4, 2022 (edited) And then I found a for sale ad for a Luxman A3600 that happened to have a pic of the 6550 conversion instructions in Kanji, which I have matched with the Hatena blog set based on the common structure and circuit references. There is one difference at the end, step 9)... 1) Change driver stage from 6240G to 6AQ8 2) Cathode resistor of driver tube (R111?) changed from 8200Ω 2W → 12kΩ 2W 3) Change constant due to power tube change (1) Add 22kΩ 1/4W between VR of input circuit and first stage input capacitor (2) Remove the cathode capacitors C103 & C104 to reduce the first stage gain (3) Change value of C113 and C114 styrol condensers from 560pF → 330pF (4) Resistance of NF circuit (C113) - R135 & R136 from 22Ω 3W → 11Ω 3W (5) Power supply circuit - R305 from 820Ω → 8.2kΩ (6) …and change R301 from 1kΩ → 2.2kΩ (7) Bias power supply circuit - add 33kΩ 1W to C307 (positive to negative) (8) ...and add 27kΩ 1/2W between preamp power connector terminal 7 and bias power supply (C307) (9) Remove decoupling resistor (R303?) 240kΩ in power supply circuit. 4) Change power tube from 8045G to KT88 5) Additional wiring in SG circuit between OPT/SG tap terminal and No.4 terminal of GT socket 6) In the case of triode connections, connect 4 points with 100Ω 1/2W between the 3rd and 4th terminals of the GT socket. Change the value of R115 and R116 from 33K-ohm 2W to 30K-ohm 2W for improvement of distortion NF resistance change R133 & R134 from 2.7kΩ 1/2W → 2.2kΩ 1/2W 7) The non-signal current of the KT88 is adjusted to 40mA, but initially it is 30mA or less and aged for 30 minutes or more, and then finally, the current is adjusted. * If you neglect this, even the KT88 may become red hot after a certain period of time. 8). In terms of performance, UL connection Triode 1kHz 50W 0.3% 1kHz 30W 0.27% 10kHz 50W 0.6% 10kHz 30W 0.4% 9) If the above distortion characteristics are worse than this, the plate resistor for the driver tube should be changed from 27kΩ → 27-47kΩ (semi-fixed). In the case of my A3600, the driver tube is still 6240G, and there was a change in the DC balance resistance (68k→39kΩ) that was not specified in the manual. It seems that it was left to the discretion of the site on a case-by-case basis Edited September 6, 2022 by Luckiestmanalive
Luckiestmanalive Posted September 6, 2022 Author Posted September 6, 2022 (edited) I found another blog with what look to be the same set of kanji instructions but typed out on the webpage so google translate had an easier job and I've revised my English version in my last post. There are three further changes I am contemplating making to my amp in line with instructions in the list above (in this order): Remove the 20pf ceramic capacitors (C114 and C115) from the circuit (I've seen pictures of A3600s without them) Remove decoupling resistor (R303) 240kΩ in power supply circuit Change the cathode resistor of driver tube (R111 and R112) from 8.2kΩ 2W → 12kΩ 2W I might try each one, in turn, and after each, monitor changes in voltages and power and, if all good, then how much difference, if any, it makes to the quality of the amp's music making. I'm also intending on buying multi-turn potentiometers to replace the 30kohm and 10kohm ones used for biasing but I have questions I will ask in a separate thread. Edited September 6, 2022 by Luckiestmanalive
RoHo Posted September 7, 2022 Posted September 7, 2022 2 hours ago, Luckiestmanalive said: I found another blog with what look to be the same set of kanji instructions but typed out on the webpage so google translate had an easier job and I've revised my English version in my last post. There are three further changes I am contemplating making to my amp in line with instructions in the list above (in this order): Remove the 20pf ceramic capacitors (C114 and C115) from the circuit (I've seen pictures of A3600s without them) Remove decoupling resistor (R303) 240kΩ in power supply circuit Change the cathode resistor of driver tube (R111 and R112) from 8.2kΩ 2W → 12kΩ 2W I might try each one, in turn, and after each, monitor changes in voltages and power and, if all good, then how much difference, if any, it makes to the quality of the amp's music making. I'm also intending on buying multi-turn potentiometers to replace the 30kohm and 10kohm ones used for biasing but I have questions I will ask in a separate thread. Looking at the schematic: 1) Ceramics are providing the negative feedback. Removing them removes the negative feedback which is a common mod. Report your findings! 2) This resistor is a bleed resistor - when the amp is turned off the charge stored in the PS caps bleeds out through it so that the PS voltage becomes zero quickly. It's there as a safety feature. Remove it and the PS will stay at HIGH VOLTAGE for much longer when the amp is TURNED OFF. Leave it in. 3) This will change the operating point of the driver circuit away from spec. IMO unlikely to improve the sound. 1
Luckiestmanalive Posted September 7, 2022 Author Posted September 7, 2022 (edited) Thanks, R - that really helps! I might purchase a couple of Mica replacements first so that if I don't like the results, I can replace them instead. Wow - OK - I will definitely leave that resistor alone! I think the change in resistor value is to improve operating point from replacing the 6240Gs with 6AQ8s. I am using 6GU7s instead (and the sound is fabulous), which might not benefit from this change in resistance. Also, I am trying to get a hold of a pair of E80CCs, which are a much closer match to the characteristics of the 6240Gs, so that is why it is last in my list. I am interested in figuring out what the load-lines look like at each stage, but especially for the driver and output stages as these are the tubes that are changed. I have found https://www.vtadiy.com/loadline-calculators/loadline-calculator/ but just need a little help inputting the right data from my amp - see my other thread. Edited September 7, 2022 by Luckiestmanalive
Luckiestmanalive Posted September 8, 2022 Author Posted September 8, 2022 (edited) Oh, and there was one other change that interested me in the power supply. There is a choke and a couple of capacitor and resistor rectifier steps. The instructions suggest changing one of the resistors (R301) from 1Kohm to 2.2Kohm (see circuit diagram below). Would this be OK to try? Edited September 8, 2022 by Luckiestmanalive
RoHo Posted September 8, 2022 Posted September 8, 2022 That resistor is part of the PS filtering system. Increasing it means the filtering will increase but you will drop more voltage - a bit more than twice as much. If you know how much current the input and drivers draw you can calculate this. So the PS voltage at the driver and input will be somewhat lower. If you have changed driver tubes this may be advantageous. Note the input stage (B3, B4) PS voltage will drop as well. The aim is usually to have the tubes operating in a linear part of the curve and maintain linearity as much as possible with signal voltage swings. Remember also, in tube circuits changes in values/voltages of less than 10% are not generally considered significant. 1
Luckiestmanalive Posted September 8, 2022 Author Posted September 8, 2022 (edited) Thanks very much, R! Here are readings taken from the measurement points on my amp vs nominal values from the circuit diagram (see pencilled figures in diagram below): B3 and B4 voltage is 257V vs 260V (new measurement 10/9 is 247V) B2 voltage is 495V vs 470V (new measurement 10/9 is 480V) B1 voltage is 520V vs 495V (new measurement 10/9 is 515V) My reading for B1 was taken from pins 3 and 8 of output tube socket. Plugging in the B3 value (257V) for V+ and resistance of 2x plate resistance value for the 6GU7s (2x 5500ohms) into the load-line webpage, it looks like the 6AQ8 is in a golden operating point. Doing the same for the 6GU7s with the 350V schematic value and 4x plate resistance value of KT120s (4x 3000ohms) also looks pretty good. If I am using the right values then I don't need to change anything! Edited September 10, 2022 by Luckiestmanalive Took new measurements 1
Luckiestmanalive Posted September 9, 2022 Author Posted September 9, 2022 (edited) Thanks, for your help, Roho - the power supply bit of the schematic (and the role of the 1Kohm and 100Kohm resistor values) make sense to me now! Edited September 9, 2022 by Luckiestmanalive
Luckiestmanalive Posted September 10, 2022 Author Posted September 10, 2022 I'm slowly ruling out the "extra" instructions from the Japanese list: 1) Change the driver tube from 6240G to 6AQ8 - don't want to use 6AQ8 as driver tube 2) Cathode resistor of driver tube (R111?) changed from 8200Ω 2W → 12kΩ 2W 3) Change constant due to power tube change (1) Add 22kΩ 1/4W between VR of input circuit and first stage input capacitor - yes, already done (2) Remove the cathode capacitors C103 & C104 to reduce the first stage gain (3) Change value of C113 and C114 styrol condensers from 560pF → 330pF - yes, already done (4) Resistance of NF circuit (C113) - R135 & R136 from 22Ω 3W → 11Ω 3W - yes, already done (5) Power supply circuit - R305 from 820Ω → 8.2kΩ - yes, already done (6) …and change R301 from 1kΩ → 2.2kΩ (7) Bias power supply circuit - add 33kΩ 1W to C307 (positive to negative) - added a 10Kohm resistor to preamp circuit to achieve same effect (8) ...and add 27kΩ 1/2W between preamp power connector terminal 7 and bias power supply (C307) (9) Remove decoupling resistor (R303?) 240kΩ in power supply circuit. - Left this current bleeding resistor in for safety's sake Roho has convinced me to leave 3.2 and 3.6 alone for a while (thanks!) - until I understand more about how they might affect the circuit because they will affect the operating points of the driver and output tubes. In the meantime, I spent the morning checking values from a lot of the resistors and capacitors in the amp. All of them were well matched and within spec except for 47uf capacitors C105 and C106, which I took readings of 74.3uf and 46.4uf. The coupling capacitor readings were all over the place so I replaced them with some Panasonic ones I received on order a while ago. I will order replacements for C105, C106 and the 0.22uf NF ceramic capacitors, C115 and C116. I'm also working on replacement bias pots and then I might be done fussing and tinkering.
Luckiestmanalive Posted September 10, 2022 Author Posted September 10, 2022 I forgot to do the readings for the four resistors attached to the output tubes yesterday so I've finished doing the readings for them, too, now. The 10ohm, 4.7Kohm and 100Kohm resistors all match very closely between them and around the nominal values but the 68Kohm resistors (blue, grey, orange) are all reading values between 13.88Kohm and 14.58Kohm. I don't know what to make of that? Those resistors are connected between pin 7 of octet terminal of the KT120s to either pin 1 or pin 3 of the balance bias pot.
Recommended Posts