Keep in mind alot of those phone W/F testing apps show the absolute RPM being off by a little, but the speed fluctuation measurements are typically dead on. This is the result I get measuring my SL1200MK2 with such an app. I cross checked it with others as well as with a tachometer.
The coreless motor on the new Technics TT models is interesting. Its also funny how some audiophiles swear they can perceive "cogging" on the "inferior" SL1200 motor. When you ask them how they evaluate this, they have no concrete answer. Fact is the cogging force on the cored Technics SL1200MK2/3/4/5 is less than the spindle bearing friction. Its literally just a couple of grams and falls within the rumble spec. That is -80 dB or more down in level. I've worked with the Neumann VMS15 cutting lathe, which is speced on paper as having worse W/F. This is a popular cutting lathe which countless records are mastered on. So far, nobody has complained about the sound of it
The coreless motor on the new Technics TT models is interesting. Its also funny how some audiophiles swear they can perceive "cogging" on the "inferior" SL1200 motor. When you ask them how they evaluate this, they have no concrete answer. Fact is the cogging force on the cored Technics SL1200MK2/3/4/5 is less than the spindle bearing friction. Its literally just a couple of grams and falls within the rumble spec. That is -80 dB or more down in level. I've worked with the Neumann VMS15 cutting lathe, which is speced on paper as having worse W/F. This is a popular cutting lathe which countless records are mastered on. So far, nobody has complained about the sound of it
Agreed. Personally, I'm not sure that cogging even has a "sound" that's distinguishable from W/F caused by anything else. In his blog, Alex Korf wrote about the relationship/difference between cogging and torque ripple (https://korfaudio.com/blog96), pointing out that the terms are often (and on the part of manufacturers, perhaps intentionally) confused. I've never seen a published "cogging force" spec, but Korf seems to agree with your point in saying that it's minimal compared to other forces affecting TTs w/ slotted DD motors. Further, he says that modulating the drive current can even compensate for it, which is something Technics has presumably been able to do since 1979.
This isn't to say Technics' new motors aren't a huge leap forward—they are—but I think the big advancement is on the main PCB. Computers have a come a long way even since the MK5. Chips 30 years ago were fast enough to process feedback in real time. According to Korf (in the post following the one I linked) they're now fast enough to "feed forward," meaning that the computer can analyze its own error data and pre-fill a table with a correction for every angular position on the next rotation.
If you go to the YouTube channel for Oswalds Mill Audio, the guy has videos where he talks (at length) about how his awesome/insane K3 table is able to serve correction data down to arc seconds, meaning the table makes 1.2m corrections per revolution. Overkill? Sure. Is there another kind of kill?
This isn't to say Technics' new motors aren't a huge leap forward—they are—but I think the big advancement is on the main PCB. Computers have a come a long way even since the MK5. Chips 30 years ago were fast enough to process feedback in real time. According to Korf (in the post following the one I linked) they're now fast enough to "feed forward," meaning that the computer can analyze its own error data and pre-fill a table with a correction for every angular position on the next rotation.
If you go to the YouTube channel for Oswalds Mill Audio, the guy has videos where he talks (at length) about how his awesome/insane K3 table is able to serve correction data down to arc seconds, meaning the table makes 1.2m corrections per revolution. Overkill? Sure. Is there another kind of kill?
Last edited:
The key thing to remember about any W/F spec IMO is the fact that they say "percent" when really what you're looking at is either a mean or median deviation from 33.3X over the course of however many sampled rotations.
This MIGHT be useful for an on-paper comparison one deck's basic performance against another, but it's not the same thing as the absolute number, length and degree of deviations we hear while listening to a record. Presumably, those momentary deviations in the pitch domain are what makes various good-sounding decks sound "different" from one another. Also, far as I know, there's only one tool capable of measuring them... 🧠
This MIGHT be useful for an on-paper comparison one deck's basic performance against another, but it's not the same thing as the absolute number, length and degree of deviations we hear while listening to a record. Presumably, those momentary deviations in the pitch domain are what makes various good-sounding decks sound "different" from one another. Also, far as I know, there's only one tool capable of measuring them... 🧠
Last edited:
Project update! (It's been a while, I know.)
Here are a few of the final drawings for the plinth.
I sent these (and additional pages with detailed measurements for every face that needs to be machined) to a pro shop a couple weeks ago. I'm now in their queue and this project should get on their CNC in May or June. This is in no way a knock against my friend who'd originally volunteered his shop for this project - he may well read this one day - it's still a long wait, but the more I thought about it, the more it made sense to have the work done locally.
Assuming everything fits, I'll likely make the full plans available somehow - for anyone else crazy enough to try re-plinthing one of these.
Here are a few of the final drawings for the plinth.
I sent these (and additional pages with detailed measurements for every face that needs to be machined) to a pro shop a couple weeks ago. I'm now in their queue and this project should get on their CNC in May or June. This is in no way a knock against my friend who'd originally volunteered his shop for this project - he may well read this one day - it's still a long wait, but the more I thought about it, the more it made sense to have the work done locally.
Assuming everything fits, I'll likely make the full plans available somehow - for anyone else crazy enough to try re-plinthing one of these.
Ah, thank you @6L6.
Now that I have a timeline for the plinth, I'm feeling motivated. But it's also a bit of a long timeline, at least a month or so, which gives me plenty of time to overthink the rest of the build. And so we come to this......
I kinda wanna give this thing a linear regulated power supply.
I know, I know. Many of you are thinking "there's NO practical reason to do that." And you're right: There's no PRACTICAL reason to do that. The only reason to do it is that I don't know how to design a point-to-point linear power supply, many of you guys do, and this website is called diyAudio.
As a starting point, I used your 6L6 Super SL PSU.
Obviously I lack an O.G. SL-1200 mains trafo, so I'm looking to grab a $30 Chinese R-core (from a seller with good feedback): 30VA with either a 30-0 or 15-0-15 secondary. Also thinking of grabbing one of those little discrete Cree Schottky bridge PCBs from Parts Connexion. Otherwise, the circuit is roughly the same: Snubber before the bridge, CRCRC with 4700uF electrolytics, LM317A, couple film caps across the output. The 100c/Mk7 wants 24v (the Mk2 wanted only 21v), so I messed around with the filter R values and regulator components in until I got +24vdc on the output.
Caveat: Chat GPT 4.0 helped me with the math, and walked me thru an LTSpice tutorial.
I would LOVE IT if some experienced people could take a look at this schematic and tell me if anything looks wrong!
(Seriously: I let an AI mess with a PSU circuit. No way in hell am I plugging it into a wall socket without an expert's opinion 😵)
A few notes are below the image.
Notes:
• D1-D4: The Cree Schottkys aren't in the spice library. 1N5819 looked similar enough.
• D6: This will probably be a 1N4002 or -07 in the real circuit. I have a few in a drawer.
• I opted to manually simulate a voltage source, rather than model the transformer. Here's how I got 34v at the bridge input:
• Technics manual says the deck consumes 8W. Ohm says if it needs 24v, it draws 330mA.
• The AI says a 30VA trafo with 115-in/30-out windings will make 34v.
(To measure output voltage in .op, I made it look like 34v dc)
Ripple also looks good:
Notes:
• Green is taken at the input of the CRCRC filter. Blue is before the regulator. Red is at the output.
• To measure .tran I switched the voltage source back to 34v @ 60Hz with no DC offset.
• R6 simulates the load on the output. Chat GPT came up with 73 ohms based on the 8W power consumption figure.
I have ideas/questions about wiring this up and connecting it to the turntable, but I'd TRULY appreciate any comments/feeback on the circuit itself first.
Thanks so much for looking.
Now that I have a timeline for the plinth, I'm feeling motivated. But it's also a bit of a long timeline, at least a month or so, which gives me plenty of time to overthink the rest of the build. And so we come to this......
I kinda wanna give this thing a linear regulated power supply.
I know, I know. Many of you are thinking "there's NO practical reason to do that." And you're right: There's no PRACTICAL reason to do that. The only reason to do it is that I don't know how to design a point-to-point linear power supply, many of you guys do, and this website is called diyAudio.
As a starting point, I used your 6L6 Super SL PSU.
Obviously I lack an O.G. SL-1200 mains trafo, so I'm looking to grab a $30 Chinese R-core (from a seller with good feedback): 30VA with either a 30-0 or 15-0-15 secondary. Also thinking of grabbing one of those little discrete Cree Schottky bridge PCBs from Parts Connexion. Otherwise, the circuit is roughly the same: Snubber before the bridge, CRCRC with 4700uF electrolytics, LM317A, couple film caps across the output. The 100c/Mk7 wants 24v (the Mk2 wanted only 21v), so I messed around with the filter R values and regulator components in until I got +24vdc on the output.
Caveat: Chat GPT 4.0 helped me with the math, and walked me thru an LTSpice tutorial.
I would LOVE IT if some experienced people could take a look at this schematic and tell me if anything looks wrong!
(Seriously: I let an AI mess with a PSU circuit. No way in hell am I plugging it into a wall socket without an expert's opinion 😵)
A few notes are below the image.
Notes:
• D1-D4: The Cree Schottkys aren't in the spice library. 1N5819 looked similar enough.
• D6: This will probably be a 1N4002 or -07 in the real circuit. I have a few in a drawer.
• I opted to manually simulate a voltage source, rather than model the transformer. Here's how I got 34v at the bridge input:
• Technics manual says the deck consumes 8W. Ohm says if it needs 24v, it draws 330mA.
• The AI says a 30VA trafo with 115-in/30-out windings will make 34v.
(To measure output voltage in .op, I made it look like 34v dc)
Ripple also looks good:
Notes:
• Green is taken at the input of the CRCRC filter. Blue is before the regulator. Red is at the output.
• To measure .tran I switched the voltage source back to 34v @ 60Hz with no DC offset.
• R6 simulates the load on the output. Chat GPT came up with 73 ohms based on the 8W power consumption figure.
I have ideas/questions about wiring this up and connecting it to the turntable, but I'd TRULY appreciate any comments/feeback on the circuit itself first.
Thanks so much for looking.
Wow, missed this one. Great project and nice find way back when you got the remains of the TT.
I've been looking for a donor TT for ages now to do what you are, the difference being that I'm a retired tight **** so it's taking me too long. I only have a pretty basic Onkyo DD (CP700M) which uses a non-PLL Matsushita drive.
Now watching with interest.
I've been looking for a donor TT for ages now to do what you are, the difference being that I'm a retired tight **** so it's taking me too long. I only have a pretty basic Onkyo DD (CP700M) which uses a non-PLL Matsushita drive.
Now watching with interest.
Great addition to the project!
To clarify, this turntable only needs 0-24VDC at the input, and merely 330mA? Ok, that’s easy. (It also seems a little small based on startup, but that’s the SP-10 in my head talking more than anything, and may be no issue at all)
Some random thoughts -
1) If you don't have something suitable already, start with a 24volt, 1A (or more) wall wart so you have something that works on hand.
2) Generally your PSU looks good. Dave brings up a good point where the filter resistors copied from my DC supply may hinder startup, they may need adjustment down IF that’s an issue, but don’t worry until it is. A decent size cap on the output of the regulator will help, of course.
3) 30VAC transformer is going to make about 40VDC at the bridge, that’s a LOT of voltage the regulator needs to drop, a smaller voltage transformer may be more appropriate. Standard regulators only need about 3V overhead to always operate properly.
4)There are lots of LM317 PCB in the world, on ebay and other places, do you have a space constraint or are you planning to make an outboard box? If using an actual transformer, it’s going to behoove you to have it outboard.
To clarify, this turntable only needs 0-24VDC at the input, and merely 330mA? Ok, that’s easy. (It also seems a little small based on startup, but that’s the SP-10 in my head talking more than anything, and may be no issue at all)
Some random thoughts -
1) If you don't have something suitable already, start with a 24volt, 1A (or more) wall wart so you have something that works on hand.
2) Generally your PSU looks good. Dave brings up a good point where the filter resistors copied from my DC supply may hinder startup, they may need adjustment down IF that’s an issue, but don’t worry until it is. A decent size cap on the output of the regulator will help, of course.
3) 30VAC transformer is going to make about 40VDC at the bridge, that’s a LOT of voltage the regulator needs to drop, a smaller voltage transformer may be more appropriate. Standard regulators only need about 3V overhead to always operate properly.
4)There are lots of LM317 PCB in the world, on ebay and other places, do you have a space constraint or are you planning to make an outboard box? If using an actual transformer, it’s going to behoove you to have it outboard.
Last edited:
There is a definite improvement in keeping any power supply as far from the phono cartridge as possible.
However, I recommend just going to an external switched power supply.
I did something similar in the past to a very beat up SL-Q2. Turned it into a battery powered luggable turntable.
I had no performance issues running it from a cheap lithium battery bank with a switching regulator. I didn't even bother with a filter for switching noise.
If you want to stick with a linear power supply, pick out a 20-22VAC transformer.
( 22V X 1.414 ) - 1.2V = 29.908V
The 1.414 represents the peak voltage of a 22V RMS AC signal. The 1.2V is worst case voltage drop across 2 diodes.
Now you have more than enough overhead for your filter and voltage regulator, without entirely too much excess. Any excess voltage will be dissipated by the voltage regulator as heat.
However, I recommend just going to an external switched power supply.
I did something similar in the past to a very beat up SL-Q2. Turned it into a battery powered luggable turntable.
I had no performance issues running it from a cheap lithium battery bank with a switching regulator. I didn't even bother with a filter for switching noise.
If you want to stick with a linear power supply, pick out a 20-22VAC transformer.
( 22V X 1.414 ) - 1.2V = 29.908V
The 1.414 represents the peak voltage of a 22V RMS AC signal. The 1.2V is worst case voltage drop across 2 diodes.
Now you have more than enough overhead for your filter and voltage regulator, without entirely too much excess. Any excess voltage will be dissipated by the voltage regulator as heat.
Last edited:
@Stuey -
Thank you and yes - DOOOO it. I'm definitely lucky to have found a donor table with the current-generation coreless/slotless/non-cogging motor. Surely another will pop up at some point - perhaps you can set up an eBay alert that emails you anytime "PARTS" or "NOT WORKING" appears in a listing with "Technics SL-1200G" "GAE" "mK7" "1500c" etc.
If an SP-10 isn't in the cards, IMO using Technics' latest, greatest motor is the next best thing... but in a high-mass, properly-damped plinth (with good arm), I think any classic PLL-enhanced Technics motor/platter might dust your Onkyo. A while back I traded DMs on instagram with a guy who did a beautiful replinth on a classic SL-1500. And here's a listing for the guts of a 1200 Mk3 for $130.
Thank you and yes - DOOOO it. I'm definitely lucky to have found a donor table with the current-generation coreless/slotless/non-cogging motor. Surely another will pop up at some point - perhaps you can set up an eBay alert that emails you anytime "PARTS" or "NOT WORKING" appears in a listing with "Technics SL-1200G" "GAE" "mK7" "1500c" etc.
If an SP-10 isn't in the cards, IMO using Technics' latest, greatest motor is the next best thing... but in a high-mass, properly-damped plinth (with good arm), I think any classic PLL-enhanced Technics motor/platter might dust your Onkyo. A while back I traded DMs on instagram with a guy who did a beautiful replinth on a classic SL-1500. And here's a listing for the guts of a 1200 Mk3 for $130.
@Dave Cawley -
Thanks so much for chiming in. To answer your specific questions:
Indeed, ADJ was connected to R4 in the circuit I "borrowed" from 6L6. So initially, I had them connected on my own schematic as well - but during the process of experimenting with the components/values in LTSpice, I ended up eliminating it. As a noob - I just know that disconnecting them made the simulation happy because it gave me almost bang-on 24v on the output. But I really don't know why R4 needs to be connected to the regulator's ADJ pin - or the potential consequences of leaving them disconnected (I'm hoping you guys will school me, so I don't burn my house down 🙂).
C1, C2 and C3 are 4700 uF. (The schematic says 0.0047 -- that's Farads, because I haven't figured out to make LTspice let me Microfarads)
Incidentally: The snubber (C4) is .22uF; (C5) is 1uF 50v; (C6) is 10 uF 50v.
As 6L6 pointed out in the post after yours, the filter resistors are there because they were there in his original circuit, I just changed the values to get the voltage I wanted on the input of the regulator. So:
R1 and R2 are 5Ω 5W . R3 (the bleeder) is 1K5 1W . R4 is 240Ω....and to get the output voltage perfectly dialed, R5 is 4k3Ω.
(I'll probably make R4 and R5 1W - maybe overkill? Also, many schematics I looked at have a Zener where I currently have R5.)
Question: When you mention startup current -- do you mean inrush?
Or should I be expecting a current spike when I start up the turntable's motor?
Thanks so much for chiming in. To answer your specific questions:
Indeed, ADJ was connected to R4 in the circuit I "borrowed" from 6L6. So initially, I had them connected on my own schematic as well - but during the process of experimenting with the components/values in LTSpice, I ended up eliminating it. As a noob - I just know that disconnecting them made the simulation happy because it gave me almost bang-on 24v on the output. But I really don't know why R4 needs to be connected to the regulator's ADJ pin - or the potential consequences of leaving them disconnected (I'm hoping you guys will school me, so I don't burn my house down 🙂).
C1, C2 and C3 are 4700 uF. (The schematic says 0.0047 -- that's Farads, because I haven't figured out to make LTspice let me Microfarads)
Incidentally: The snubber (C4) is .22uF; (C5) is 1uF 50v; (C6) is 10 uF 50v.
As 6L6 pointed out in the post after yours, the filter resistors are there because they were there in his original circuit, I just changed the values to get the voltage I wanted on the input of the regulator. So:
R1 and R2 are 5Ω 5W . R3 (the bleeder) is 1K5 1W . R4 is 240Ω....and to get the output voltage perfectly dialed, R5 is 4k3Ω.
(I'll probably make R4 and R5 1W - maybe overkill? Also, many schematics I looked at have a Zener where I currently have R5.)
Question: When you mention startup current -- do you mean inrush?
Or should I be expecting a current spike when I start up the turntable's motor?
He is talking about motor start current.
The thing that strikes me about the AI circuit, which I didnt see on the first look, is the set resistors (R4 R4 in your schematic) are not connected properly.
This is the way that works -
Of course, the adjust resistor doesn't need to be a potentiometer wired as variable resistor as shown in this schematic, it can be a fixed resistor.
But notice that R1 is connected from OUT to ADJ, and R2 is from ADJ to GND.
Vout is 1.25 x (1+ R2/R1)
Do yourself a favor and use 240R for R1, it's on the datasheet example for a reason. 🙂
Here's a nifty calculator if you don't want to do the math - https://circuitdigest.com/calculators/lm317-resistor-voltage-calculator
I've been very happy with fixed regulators lately, somebody pointed out that the Onsemi MC78xx /79xx series regulators are notably better performance than the jellybean copies of the Ti part, which is handy considering they aren't really that much more money, just a few cents. There's absolutely nothing wrong with the LM317 and no reason not to use it... the fixed regulators only need capacitors and that makes things simpler. Appropriate part for this application - https://www.digikey.com/short/cz85159b
The thing that strikes me about the AI circuit, which I didnt see on the first look, is the set resistors (R4 R4 in your schematic) are not connected properly.
This is the way that works -
Of course, the adjust resistor doesn't need to be a potentiometer wired as variable resistor as shown in this schematic, it can be a fixed resistor.
But notice that R1 is connected from OUT to ADJ, and R2 is from ADJ to GND.
Vout is 1.25 x (1+ R2/R1)
Do yourself a favor and use 240R for R1, it's on the datasheet example for a reason. 🙂
Here's a nifty calculator if you don't want to do the math - https://circuitdigest.com/calculators/lm317-resistor-voltage-calculator
I've been very happy with fixed regulators lately, somebody pointed out that the Onsemi MC78xx /79xx series regulators are notably better performance than the jellybean copies of the Ti part, which is handy considering they aren't really that much more money, just a few cents. There's absolutely nothing wrong with the LM317 and no reason not to use it... the fixed regulators only need capacitors and that makes things simpler. Appropriate part for this application - https://www.digikey.com/short/cz85159b
I think 6L6 has given you the answers you asked me ! And I agree that a proper regulator from a proper distributor is the only way to go. Cheap 317's do not perform as well. And the 7824 is a very good out of the box regulator (assuming it's a good one as 6L6 says). BTW Timestep (my company) used the "National Semiconductor LM317HVT" exclusively, and yes a 240ohm.
I have noticed a growing tendency on this forum to rely on simulation instead of proper measurements, and I think here lies the R4 issue. Simulators do not take into account of mistakes, board layout and real world (Chinese) components. Hence IMHO it's better to design from the manufactures data and then build one and rigorously test it on and off load and start-up using a good quality Scope and DVM. Use a resistive load to start with not the turntable until it's all good, then when you are sure connect the turntable and repeat the measurements.
Had you thought about the SL-100c bearing ?
YMMV Dave
I have noticed a growing tendency on this forum to rely on simulation instead of proper measurements, and I think here lies the R4 issue. Simulators do not take into account of mistakes, board layout and real world (Chinese) components. Hence IMHO it's better to design from the manufactures data and then build one and rigorously test it on and off load and start-up using a good quality Scope and DVM. Use a resistive load to start with not the turntable until it's all good, then when you are sure connect the turntable and repeat the measurements.
Had you thought about the SL-100c bearing ?
YMMV Dave
Thanks, Dave. 6L6 has been massively helpful from the start, but I really appreciate everyone who has taken time to answer questions/offer ideas about this project. Takes a village as they say.
Three things I now realize are related:
1. About startup current - Dave, I'd actually would love it if you could elaborate. Initially I did think you meant inrush, but thanks to the clarification now realize you're saying the motor needs extra current to "wake up," and the filter might drop too many volts to supply it. But if that's the case, is it just a matter of knowing how much current it will draw and for how long, and then making sure the cap across the output can deal with it? Or do the filter+regulator affect that cap's response in some way that I just don't understand?
2. I've decided to go with that Onsemi fixed regulator (and, at Dave's advice, I read the datasheet). So, speaking of that output cap:
My assumption:
Cin is the voltage reserve that prevents the regulator from dropping out.
Cout is the voltage reserve that ensures stability across the actual load.
The datasheet specs .33uF on the input, but wouldn't that really depend on how much headroom you're giving the regulator? Also, it only gives a minimum for the output, and for a DC motor that dissipates 8W to run at an ultra-stable RPM (and spikes to...what? 16W at startup?) seems like that minimum is gonna be a LOT higher than .1? I'm guessing there's a formula for Cout.... and I haven't fully digested that datasheet, so maybe it's on there. Also, any insight on how to measure that startup current would be appreciated (I do have the table's stock SPMPS, so taking a real-world measurement should be no problem. Should I just set the multimeter to amps and clip probes across the input?)
3. 6L6 and Akimmet both said 30VA, 30v was too much trafo. (AI helped me spec it...I know, caveat emptor). Initially I just told it (1) no hum, (2) minimal stray flux, and (3) give the regulator plenty o' headroom. Using a 30VA trafo to run a 330mA load takes care of 1 and 2... But that caused the AI to go WAY overboard on the headroom request. (What I neglected to tell it was "please don't melt the regulator.")
So... as you guys pointed out, calculating how many volts of headroom I have on my regulator input is as simple as figuring out the Vpeak after the rectifier and subtracting the voltage drop across the filter. But something I ALSO learned yesterday is if I want to know how many volts of headroom are actually at the regulator, I need to calculate Vpeak DC based on the actual load on the trafo secondary, and an 8W device across a 30 VA transformer is using less than 30% of its capacity and therefore Vpeak is MUCH HIGHER than what you'd get by multiplying the rated secondary voltage by 1.414. (For Lolz - ChatGPT gave the reg. like 14.5 V of headroom).
So... I'm thinking about this 20VA trafo with a 22v output winding.
(330mA across 20VA is more like 40% loaded, so Vpeak is still higher than it would otherwise be. I did this in Spice but as Dave kindly pointed out, I need to breadboard this.) Wow, I've learned a lot this week. But PLEASE FEEL FREE TO POKE ANY HOLES YOU'D LIKE IN ANY OF THIS!!
Again, thank you guys. This community is AWESOME.
Three things I now realize are related:
1. About startup current - Dave, I'd actually would love it if you could elaborate. Initially I did think you meant inrush, but thanks to the clarification now realize you're saying the motor needs extra current to "wake up," and the filter might drop too many volts to supply it. But if that's the case, is it just a matter of knowing how much current it will draw and for how long, and then making sure the cap across the output can deal with it? Or do the filter+regulator affect that cap's response in some way that I just don't understand?
2. I've decided to go with that Onsemi fixed regulator (and, at Dave's advice, I read the datasheet). So, speaking of that output cap:
My assumption:
Cin is the voltage reserve that prevents the regulator from dropping out.
Cout is the voltage reserve that ensures stability across the actual load.
The datasheet specs .33uF on the input, but wouldn't that really depend on how much headroom you're giving the regulator? Also, it only gives a minimum for the output, and for a DC motor that dissipates 8W to run at an ultra-stable RPM (and spikes to...what? 16W at startup?) seems like that minimum is gonna be a LOT higher than .1? I'm guessing there's a formula for Cout.... and I haven't fully digested that datasheet, so maybe it's on there. Also, any insight on how to measure that startup current would be appreciated (I do have the table's stock SPMPS, so taking a real-world measurement should be no problem. Should I just set the multimeter to amps and clip probes across the input?)
3. 6L6 and Akimmet both said 30VA, 30v was too much trafo. (AI helped me spec it...I know, caveat emptor). Initially I just told it (1) no hum, (2) minimal stray flux, and (3) give the regulator plenty o' headroom. Using a 30VA trafo to run a 330mA load takes care of 1 and 2... But that caused the AI to go WAY overboard on the headroom request. (What I neglected to tell it was "please don't melt the regulator.")
So... as you guys pointed out, calculating how many volts of headroom I have on my regulator input is as simple as figuring out the Vpeak after the rectifier and subtracting the voltage drop across the filter. But something I ALSO learned yesterday is if I want to know how many volts of headroom are actually at the regulator, I need to calculate Vpeak DC based on the actual load on the trafo secondary, and an 8W device across a 30 VA transformer is using less than 30% of its capacity and therefore Vpeak is MUCH HIGHER than what you'd get by multiplying the rated secondary voltage by 1.414. (For Lolz - ChatGPT gave the reg. like 14.5 V of headroom).
So... I'm thinking about this 20VA trafo with a 22v output winding.
(330mA across 20VA is more like 40% loaded, so Vpeak is still higher than it would otherwise be. I did this in Spice but as Dave kindly pointed out, I need to breadboard this.) Wow, I've learned a lot this week. But PLEASE FEEL FREE TO POKE ANY HOLES YOU'D LIKE IN ANY OF THIS!!
Again, thank you guys. This community is AWESOME.
PS -- it's really small in my photo, but here's the voltages at each stage:
32.7 on the bridge input
32.5 on the first cap of the filter
32.3 a the regulator input (8.3v headroom, 7.7v below max)
And the current BOM:
Mains - 20VA, 0-22v secondary - eBay China
Snubber cap - Kemet R82EC3220DQ70J .22uF 63V
Discrete Schottky diode bridge - Wolfspeed CSD01060A
Filter Caps - 4700uF 63V
Filter Resistors - 5R 5W wirewound
Bleeder - 1K5 2W metal film
Fixed LDO Regulator - Onsemi MC7824ACTG
Regulator I/O caps:
C1 - probably .33uF (film?)
C2 - TBD
Any help with this is appreciated. Opinions on the trafo choice are most dire, I want to try and get it here before China's blowback tarriffs kick in.
PS - anyone want to sell me an O-scope?
32.7 on the bridge input
32.5 on the first cap of the filter
32.3 a the regulator input (8.3v headroom, 7.7v below max)
And the current BOM:
Mains - 20VA, 0-22v secondary - eBay China
Snubber cap - Kemet R82EC3220DQ70J .22uF 63V
Discrete Schottky diode bridge - Wolfspeed CSD01060A
Filter Caps - 4700uF 63V
Filter Resistors - 5R 5W wirewound
Bleeder - 1K5 2W metal film
Fixed LDO Regulator - Onsemi MC7824ACTG
Regulator I/O caps:
C1 - probably .33uF (film?)
C2 - TBD
Any help with this is appreciated. Opinions on the trafo choice are most dire, I want to try and get it here before China's blowback tarriffs kick in.
PS - anyone want to sell me an O-scope?
Last edited:
Buy a Rigol DHO802 right now. They are absolutely killer for $339.
For the C output, use 220uF or more. One of the nice things about the 78xx regulations is they don't really care if you put big caps downstream, so you can have more filtering as well as a bit of help on startup.
For the C output, use 220uF or more. One of the nice things about the 78xx regulations is they don't really care if you put big caps downstream, so you can have more filtering as well as a bit of help on startup.
I agree with 6L6, but I would use a better transformer. Cheap Chinese transformers are just that. You can get encapsulated, in a shielding can and an electrostatic inter-winding band. I would look at Mouser first and decide on your budget. 30VA would be my choice.
I do find the Rigol scope more difficult to use than my Keysight, the EDUX1052A is their cheapest (in the long term you may want 4 channels though)
The output capacitor for start-up/inrush, you could use one of your 4,700uF on the output as well.
IMHO, I do nothing without a scope and DVM.
YMMV Dave
I do find the Rigol scope more difficult to use than my Keysight, the EDUX1052A is their cheapest (in the long term you may want 4 channels though)
The output capacitor for start-up/inrush, you could use one of your 4,700uF on the output as well.
IMHO, I do nothing without a scope and DVM.
YMMV Dave
Last edited:
Thanks for all the advice, guys.
Digikey has a 25VA Toroid with a 24V (series) secondary from Triad. Medical grade, flux band + static shield. Under $40 bucks. If I have the math right, my 8W nominal load is about 1/3 of that trafo's capacity, which ends up giving me ~32v peak DC after the bridge.
Still more than enough headroom after the filter, only a couple watts for the regulator to dissipate. (And I'll put a good sink on the regulator. Also thinking I'll add a thermistor after the fuse on the AC side.) Extra 4700uF electrolytic is in my cart for the output - going to get this (and a scope) ordered.
Dave, one more thing - you asked if I've thought about the 100c bearing.
I have thought about it—to the extent that I realize far better bearings exist. (And also noted that rebuilding the bearing/new thrust pad is the first thing almost used SP-10 buyer does.) My bearing basically has zero miles on it... and the thing about all the new Technics motors (100c all the way up to SP-1000) is that the bearing's machined housing is directly coupled to the motor itself, so using a different one would be a pretty major engineering task.
All that said, I'm curious what your thoughts were when you asked.
Digikey has a 25VA Toroid with a 24V (series) secondary from Triad. Medical grade, flux band + static shield. Under $40 bucks. If I have the math right, my 8W nominal load is about 1/3 of that trafo's capacity, which ends up giving me ~32v peak DC after the bridge.
Still more than enough headroom after the filter, only a couple watts for the regulator to dissipate. (And I'll put a good sink on the regulator. Also thinking I'll add a thermistor after the fuse on the AC side.) Extra 4700uF electrolytic is in my cart for the output - going to get this (and a scope) ordered.
Dave, one more thing - you asked if I've thought about the 100c bearing.
I have thought about it—to the extent that I realize far better bearings exist. (And also noted that rebuilding the bearing/new thrust pad is the first thing almost used SP-10 buyer does.) My bearing basically has zero miles on it... and the thing about all the new Technics motors (100c all the way up to SP-1000) is that the bearing's machined housing is directly coupled to the motor itself, so using a different one would be a pretty major engineering task.
All that said, I'm curious what your thoughts were when you asked.
Because all the SP-10 bearings are at least 40yr old... Changing the bearing in an integrated assembly like yours would be not worth the effort at all, and would likely bring no benefit.