I don't fault the actual circuit design of this reg, but they advocate using a wire nut along with these solderless wire insert terminals, which I believe is just a lousy way to connect something so critical. The ground wire comes pre-installed into the barrier strip, but I should have checked it was tight. KAB even says to verify outout voltage despite also stating that every reg is tested upon assembly - based on the output voltage being off at first, I have a hard time believing that. This little reg costs $50. Thats a somewhat fair price for convenience, but when your selling $5 worth of parts with minimal assembly labor and charge 10 times that much, you should be making sure it works properly and all connections are solid. I did my due diligence first by testing the reg and still had to correct the output voltage.
Unfortunately I was too hasty, wanting to test the main board out before final assembly, so when I fired it up with the platter in place (to be safe), it almost immediately went into the washing machine of death mode, with the platter turning randomly forwards and backwards. This almost always means the AN6680 is toast, but the AN6675 got scorching hot just in the few seconds the power was applied. These are unforgiving ICs to over-voltage and frequently fail when the internal reg circuit develops intermittent solder joints over the years, especially Q201 - this transistor runs very hot and should be heat sinked.
Why they would cater to people who didn't know basic soldering skills for such a delicate application is beyond me. I would think anyone who tears into their expensive vintage Technics 1200 should know how to do basic soldering work.
This pic below is from the KAB site, showing the ground connection with wire nut.. absolute fail IMO. Again, not knocking the circuit design, just the solderless terminals and wire nut for the CRITICAL ground connection.
Unfortunately I was too hasty, wanting to test the main board out before final assembly, so when I fired it up with the platter in place (to be safe), it almost immediately went into the washing machine of death mode, with the platter turning randomly forwards and backwards. This almost always means the AN6680 is toast, but the AN6675 got scorching hot just in the few seconds the power was applied. These are unforgiving ICs to over-voltage and frequently fail when the internal reg circuit develops intermittent solder joints over the years, especially Q201 - this transistor runs very hot and should be heat sinked.
Why they would cater to people who didn't know basic soldering skills for such a delicate application is beyond me. I would think anyone who tears into their expensive vintage Technics 1200 should know how to do basic soldering work.
This pic below is from the KAB site, showing the ground connection with wire nut.. absolute fail IMO. Again, not knocking the circuit design, just the solderless terminals and wire nut for the CRITICAL ground connection.
Attachments
Last edited:
Not got this. On "other" PSU's if you don't connect the ground there is no power to destroy anything ?
@Dave Cawley Well this last picture is from KAB's site, not from my install. Note the wire going through the hole without a grommet. Thats really sloppy..
Yes, there's no ground connected in that picture to the motor PCB. That's a bad thing on a linear reg, with the adjustment pin floating open in the air without a ground to reference. That makes the reg output V go high, close to the input V minus the reg dropout voltage.
Yes, there's no ground connected in that picture to the motor PCB. That's a bad thing on a linear reg, with the adjustment pin floating open in the air without a ground to reference. That makes the reg output V go high, close to the input V minus the reg dropout voltage.
Those spare boards are becoming more expensive and its hard to be sure they weren't subject to previos abuse from former owners. They're fairly reliable in stock form as long as the solder connections around the internal reg circuit are all good, you can likely get away with just re-capping the board and re-flow the majority of solder joints.
The 5.1V zener diode in the main reg circuit can become flaky over the years. Usually the reverse drop V goes down but sometimes it can go open, which is obviously a much worse mode of failure. That spells bad news for the motor control ICs, especially the AN6680 and 9V sub reg Q201 fed by the main 21V reg. The AN6675 drive output chip is a little more resilient, but it gets the full 21V reg output, so its very important the entire reg circuit works properly.
There is another 5.1V zener by the AN6682 VCO IC, but so far over the 35+ years of working on these 1200 family of TTs, I've only seen that diode go down in reverse drop voltage when it fails, usually below 3V or so. That won't however hurt the other main control ICs.
@Dave Cawley Yes, I understand now what you're saying.
There's obviously the little tantalum cap on the reg input side to guarantee circuit stability. Most people who know about the risks of using this type of cap will avoid using them in applications where more durability is required. Tantalums are known to be unreliable, but they have a high DF which is good for suppressing regulator instability. I'd rather see a decent quality (non-low ESR electrolytic in that location).
A larger zener at the main reg input fed by a low value series resistor would be cheap insurance, absorbing excess voltage should the main reg input go high. That may add too much cost.
There's obviously the little tantalum cap on the reg input side to guarantee circuit stability. Most people who know about the risks of using this type of cap will avoid using them in applications where more durability is required. Tantalums are known to be unreliable, but they have a high DF which is good for suppressing regulator instability. I'd rather see a decent quality (non-low ESR electrolytic in that location).
A larger zener at the main reg input fed by a low value series resistor would be cheap insurance, absorbing excess voltage should the main reg input go high. That may add too much cost.
@Netlist I totally agree. I have a few consumable parts left for these 1200s ie. cueing lamps, pitch faders, etc but the larger main control ICs have been discontinued as you know. I have one spare working board but it came out of a unit with a seized platter. Therfore the motor was heavily loaded all the time and this would have been murder on the drive IC. Used parts for these units are highly suspect due to this, plus the thermal abuse from unskilled techs removing the parts. This is why I'd rather buy a complete used board. Problem is most of them are now 40+ years old with leaky caps, discolored and/or overheated traces, etc. The PCB substrate is usually pressed paper mixed with phenolic resin. Once this stuff overheats and darkens, it starts to get brittle, becoming conductive. The traces close to components will start to lift, creating other new unexpected issues.
Some surgery on the motor PCB later and now I at least have all of the voltages restored. The AN6675 motor drive looks to be ok and Q201 is replaced. Just looks like the AN6680 frequency generator is bad. It got cooked from Q201 failing.
Well, I was too hasty in clearing the AN6675 of fault. If you lift pin 22, the motor should not be able to start. In my case the motor starts spinning erratically upon powering up the TT and wont respond to the start/stop switch.
Some success with this problem. I had an old board from an SL1600 which uses the same drive and control ICs, so I swapped the AN6675 on the 1200 board and it fixed the problem.
However, testing W/F revealed it to be over 0.1% RMS, so about 0.12% peak. So then I removed the motor armature and installed another one. Using the spindle to center the whole thing, it was reinstalled and W/F tested to be around 0.04% peak. That's an excellent figure and it can be reduced even more with some careful centering of the magnet asy.
However, testing W/F revealed it to be over 0.1% RMS, so about 0.12% peak. So then I removed the motor armature and installed another one. Using the spindle to center the whole thing, it was reinstalled and W/F tested to be around 0.04% peak. That's an excellent figure and it can be reduced even more with some careful centering of the magnet asy.
Attachments
Less power could improve life time, see here SL-Q2 vs. SL1200/10:
Calvins audio page Technics (about 2/3 down)
12 V battery supply wouldn't vibrate like a transformer.
It would be great if you could make a list of the does and don'ts to keep the SL1200/10 alive (protection for NLA ICs, shortening wires, etc.)
Calvins audio page Technics (about 2/3 down)
12 V battery supply wouldn't vibrate like a transformer.
It would be great if you could make a list of the does and don'ts to keep the SL1200/10 alive (protection for NLA ICs, shortening wires, etc.)
@alujoe2 Its sort of a double edge sword with powering the 1200s motor. Less voltage on the AN6675 will help alot, but that reduces transient response of the drive output.
The RC filter zobels on the drive coils are designed to reduce EMF. They're sized to soften the leading edge of each drive pulse by reducing the resonant Q formed by the armature coils. The pulses can be further softened to improve flutter and cogging, but they're already on the edge of being too soft.
The cogging on 1200s isn't remotely as bad as people tend to make it sound. Thats mainly due to the fact that the motor has a 3 phase drive, plus the drive reference FB frequency is 50 kHz, far outside the range of human hearing. The platter mass will smooth out any residual peaks of drive pulses.
Further reducing drive current to the motor coils will increase the drive pulse rise time, in return reducing stress on the AN6675. This will reduce motor torque, but thats only a potential issue if the TT is used for heavy DJing.
I've experimented with .47 ohm resistors on each group of coils. The 12 drive coils are configured in 3 groups of 4 coils. The resistors need to be placed between the AN6675 drive output pins and RC zobels of each coil group. It requires cutting traces unless you want to solder the RC directly onto the coil drive pins on the armature, which is the easier option IMO.
The next way to improve AN6675 drive IC reliability is by installing a larger heat sink. This reduces the chip's operating temp but it doesn't make that much of a difference in temperature. It is however the easiest mod to perform, so its very much worth doing. A larger low profile CPU heatsink is easy to modify by drilling and tapping the 2 mounting holes. It can be easily marked for drilling by using the old heatsink as a guide. You'll need to clear the capacitors next to the AN6675. Otherwise its very simple and straightforward. A higher spec heatsink paste (ie. Artic Silver, etc) on the IC can help with cooling as well. Most of the time, just refreshing the heatsink compound makes a difference in cooling, as many of these 1200s are already over 40 yrs old by now. You can be assured the heatsink paste will be dried out by now.
Resoldering the components on the main PCB making up the voltage reg circuit help as well. If some of these parts develop bad connections, it can cause the reg's output voltage to spike, harming the AN6675 and making Q201 run warmer. Q201 doesn't have a heat sink but runs very warm, so an additional heatsink on it would improve reliability. Most of the time Q201 already has intermittent solder connections on it due to heat and vibration. The solder also breaks down over time, mainly from Q201 running hot. If your TT sometimes starts running fast, slow, backwards by itself, Q201 is the first suspect, as it forms the 9.4V reg circuit with the AN6680 control IC.
Bearing drag is also something which will increase AN6675 operating temp. Heavy platter mats and record clamps also influence this. I've seen alot of 1200s with bad bearings and they will keep going up to a point without much warning. Thats why its important to check and lubricate the bearing periodically, especially the thrust bearing pad underneath, where you can't get to. You must remove the drive PCB to access it, so it often gets neglected. This also raises bearing rumble and vibration, so its important to check every so often.
I really like the KAB thrust bearing plate to replace the stock bearing pad. It also makes a huge difference in rumble noise, so its a very worthwhile upgrade. When you do replace one of these, its important to polish the spindle bottom tip before installing it. I chuck the spindle into a drill and use 3M medium polishing compound followed by a mirror polish using plain printer paper pushed against the bearing tip surface.
Lastly, while the LM317 reg mod does improve the DC supply noise floor, it also increases the current capability to the AN6675. This can potentially hurt the IC long term with high motor load, so factor that in when changing over to an external LM317 reg. The stock reg is designed to be gentle on the AN6675, so IMO its not always a good idea to replace the stock reg.
Re-capping the main PCB can also improve reliability, as some of the caps will have deteriorated to the point they've dried up internally and changed values. In extreme cases they can also leak, eating up copper traces on the PCB. This is common on very old 1200s which have run very warm, having a hard life under heavy use. I prefer using Panasonic EB series capacitors as upgraded replacements. They're usually rated at 105 deg C, so they're very reliable and the price isn't too bad either.
Lastly, I highly recommend purchasing a backup motor PCB for parts or replacement should your TT suffer from bad ICs. They can be had for under $150 for a guaranteed working PCB, which should be recapped and resoldered before use. I found one for $70 on bleed-bay, so they definitely are out there for relatively little money. Look for unrepaired boards, otherwise its a high likelihood they won't work right and have bad ICs. Most of the time repairs are simple. These 1200s are very reliable, even being as old as they are now. I wish other things were this well engineered in the audio world.
I should start another thread about some of these topics. In the mean time I hope these pointers help.
The RC filter zobels on the drive coils are designed to reduce EMF. They're sized to soften the leading edge of each drive pulse by reducing the resonant Q formed by the armature coils. The pulses can be further softened to improve flutter and cogging, but they're already on the edge of being too soft.
The cogging on 1200s isn't remotely as bad as people tend to make it sound. Thats mainly due to the fact that the motor has a 3 phase drive, plus the drive reference FB frequency is 50 kHz, far outside the range of human hearing. The platter mass will smooth out any residual peaks of drive pulses.
Further reducing drive current to the motor coils will increase the drive pulse rise time, in return reducing stress on the AN6675. This will reduce motor torque, but thats only a potential issue if the TT is used for heavy DJing.
I've experimented with .47 ohm resistors on each group of coils. The 12 drive coils are configured in 3 groups of 4 coils. The resistors need to be placed between the AN6675 drive output pins and RC zobels of each coil group. It requires cutting traces unless you want to solder the RC directly onto the coil drive pins on the armature, which is the easier option IMO.
The next way to improve AN6675 drive IC reliability is by installing a larger heat sink. This reduces the chip's operating temp but it doesn't make that much of a difference in temperature. It is however the easiest mod to perform, so its very much worth doing. A larger low profile CPU heatsink is easy to modify by drilling and tapping the 2 mounting holes. It can be easily marked for drilling by using the old heatsink as a guide. You'll need to clear the capacitors next to the AN6675. Otherwise its very simple and straightforward. A higher spec heatsink paste (ie. Artic Silver, etc) on the IC can help with cooling as well. Most of the time, just refreshing the heatsink compound makes a difference in cooling, as many of these 1200s are already over 40 yrs old by now. You can be assured the heatsink paste will be dried out by now.
Resoldering the components on the main PCB making up the voltage reg circuit help as well. If some of these parts develop bad connections, it can cause the reg's output voltage to spike, harming the AN6675 and making Q201 run warmer. Q201 doesn't have a heat sink but runs very warm, so an additional heatsink on it would improve reliability. Most of the time Q201 already has intermittent solder connections on it due to heat and vibration. The solder also breaks down over time, mainly from Q201 running hot. If your TT sometimes starts running fast, slow, backwards by itself, Q201 is the first suspect, as it forms the 9.4V reg circuit with the AN6680 control IC.
Bearing drag is also something which will increase AN6675 operating temp. Heavy platter mats and record clamps also influence this. I've seen alot of 1200s with bad bearings and they will keep going up to a point without much warning. Thats why its important to check and lubricate the bearing periodically, especially the thrust bearing pad underneath, where you can't get to. You must remove the drive PCB to access it, so it often gets neglected. This also raises bearing rumble and vibration, so its important to check every so often.
I really like the KAB thrust bearing plate to replace the stock bearing pad. It also makes a huge difference in rumble noise, so its a very worthwhile upgrade. When you do replace one of these, its important to polish the spindle bottom tip before installing it. I chuck the spindle into a drill and use 3M medium polishing compound followed by a mirror polish using plain printer paper pushed against the bearing tip surface.
Lastly, while the LM317 reg mod does improve the DC supply noise floor, it also increases the current capability to the AN6675. This can potentially hurt the IC long term with high motor load, so factor that in when changing over to an external LM317 reg. The stock reg is designed to be gentle on the AN6675, so IMO its not always a good idea to replace the stock reg.
Re-capping the main PCB can also improve reliability, as some of the caps will have deteriorated to the point they've dried up internally and changed values. In extreme cases they can also leak, eating up copper traces on the PCB. This is common on very old 1200s which have run very warm, having a hard life under heavy use. I prefer using Panasonic EB series capacitors as upgraded replacements. They're usually rated at 105 deg C, so they're very reliable and the price isn't too bad either.
Lastly, I highly recommend purchasing a backup motor PCB for parts or replacement should your TT suffer from bad ICs. They can be had for under $150 for a guaranteed working PCB, which should be recapped and resoldered before use. I found one for $70 on bleed-bay, so they definitely are out there for relatively little money. Look for unrepaired boards, otherwise its a high likelihood they won't work right and have bad ICs. Most of the time repairs are simple. These 1200s are very reliable, even being as old as they are now. I wish other things were this well engineered in the audio world.
I should start another thread about some of these topics. In the mean time I hope these pointers help.