1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Simple DC motor control

Discussion in 'd.i.y.' started by martin clark, Dec 22, 2009.

  1. martin clark

    martin clark pinko bodger

    Following several PMs and most recently Si's Origin live thread here's an idea for a simple controller for DC turntable motors, in particular low-voltage DC motors which seem to be popular currently. These offer fairly low vibration and low 'cogging' but also brings real uncertainty in repeatability (speed control from one day to another) and comparatively poor control response to the usual drives it is paired with - which are inevitably constant-voltage devices i. e. plain old voltage regulators.

    One way used to run small motors under load accurately at low speed is to use a measure of current-feedback - basically, a small resistor in the 0v side of the motor return, and sense the voltage across this. It compensates for the armature's own resistance and seems to work really, really well (esp for minimal cogging at very low speed). This is not a new or original idea by any means - Ferranti, Philips and others used simple two and three-transistor variants for cheap capstan motors way back in the 60s, and the principle is older still. I used it for a diy cassette transport when a student and found it better than a 317...

    Anyway, various mouldy old things were dusted-off and this knocked-up to help Si (Chopsaw) drive a OL motor. Here's a circut to play with:

    [​IMG](- don't build to these values, there are updates below)

    From the voltage reference an opamp drives a pass transistor. Opamp is set-up as inverting gain of 1x, so the output voltage is compared with reference and motor voltage adjusted accordingly. A basic discrete voltage reg in other words.

    The current-sensing bit is actually +ve feedback. If the motor is loaded, the current through it increases, which is sensed by the opamps +ve input from Rs and *added* to the voltage reg output, driving the motor harder so the current falls... R3, here '3k00' is there to get the scaling of these two feedback loops correct. In other words it controls damping in the system. As a result this resistor must be calculated for each motor/instance, based on the measured armature DC resistance and the values chosen for the two feedback resistors (above, both 10k). To make that easier here is a simple excel spreadsheet.

    It may be better to use a fixed resistor plus a series trimpot for the damping setting to allow a bit of adjustment around this figure. If its far out, the motor can actually speed-up under load, or slow too easily (like normal DC reg supply) - you'll hear 'wow'. By playing around you'll find that your chosen motor feedback resistance must always be less than the motors armature reistance, or the control loop will be unstable. This example uses 4R7 for 12ohm armature (measured across the motor terminals)

    None of the circuit parts are critical; make life as hard as you wish. The opamp can be anything slow and stable, NE5534 is a great place to start. Play with decoupling arangements as you like.

    Because the main gain loop is inverting (and current-sense voltage is very close to 0v for small motors as low speed) this uses a split- rail supply. 7812/7912 quite good enough for supplies.

    The pass transistor only needs to be NPN and allow an Ic of 500mA or more (in case the motor stalls). The NPN collector runs off the +ve rail, and could be dissipating a watt or two at stall. A TO220-cased part should be OK without heatsink.

    Running off the -ve rail is the voltage reference. The 2k2 resistor feeds this about 3mA off 12v, so a 6.9v zener would be ok, an LM329 great if overkill. It's followed by a bit of filtering anyway :1K / 100uF would roll-off everything over 1.6hz - sheer brute force, really, given the reference is run off a regulated supply!

    The speed-set is adjusted with trimpots. The diagram shows 10-turn parts but 22turn pots would offer finer control. You can fudge the zener value or range-set resistors as required to suit you motor (these values were set-up to suit an Origin Live one...).

    Note that the filter and the trimpot are a voltage divider, so you'll see about 5v at the top of the trimpot. That's the max this can apply to the motor - if you want more, fiddle the resistors (R1/ R2 in the spreadsheet) to give the opamp some gain (increase R1 a bit over R2).

    Adding a cap across R2 will reduce AC gain and probably improve performance a bit. Try 1-10uF

    None of the caps matter, can use big cheap electrolytics for the filtering and bulk decoupling - just watch for polarity.

    And that's about it, really.



    Update: 25th Jan

    Well since there are PCBs going out (post #261) here's a quick update and bug-fix:
    1) the spreadsheet mentioned above has been fixed a little bit, 'save as' and have a play with the yellow boxes.

    2) Here's the correlation of sketch above and the parts on the circuit boards:

    [​IMG]
    - I'll update with with labelled photos of the boards and soon.

    3) Build & tweaks:
    • The PCB does not include for the commutation noise filter suggested by Werner that helps so much . This is easy to add 'freehand'
    • In the spreadsheet update it simplifies things to run the voltage reference side at gain of -1 (R43=R44); if you need more voltage output to get your motor to the right speed, try using wire links in place of R37, R40, R41, R42, and 10K multiturn pots for the presets. If that's still not enough - rather unlikely - swap the zener for 8v2 and try again.
    • R38 can usefully be reduced, to 500ohms or so. Definitely do this if you use a zener >6.9v.
    • You can use 7815/7915 regs instead if you want more voltage headroom for strange motors / higher voltage zeners.
    • R43, R44 values are now increased to 47Kohms. This minimises errors due to the source impedance of the voltage references, which tend to over-compensate the motor by effectively making R45 up to 7% too small. More detail is given in posts #337 and #339
    • If you get speed variation or audible wow, increase your R45 in small increments - start with +5%
    • Other minor items we can sweep-up in thread / [updates to be added].

    4) GSC Board BOM (thanks Paul!)
    [updated 30th Jan 2010]
    (PCB ref) (part value) (Farnell reference)

    U10 TLO71 SINGLE J-FET OP-AMP FEC 145-9699 ( or NE5534 )
    U12 7812 REGULATOR FEC 108-7091
    U13 7912 REGUALTOR FEC 109-5391
    TR9 BD237 MEDIUM POWER TRANSISTOR FEC 994-6560
    C18; C21-C26 100NF 50V 0603 X7R CAPS FOC supplied with the bare PCB’s
    C19 100UF 25V LOW ESR RADIAL CAP FEC 121-9466
    C20 NOT FITTED but supplied with bare PCB if you choose to fit it.
    C27/C28 10UF 25V TANT BEAD FEC 970-8448
    C29/C30 330UF 63V LOW ESR RADIAL CAP FEC 969-2983
    C31 10UF 25V RADIAL CAP FEC 812-6380
    D21-D24 1N4002 RECTIFIER DIODES 200PIV FEC 146-7453
    DZ1 BZX85 C6V8 ZENER DIODE FEC 984-4279
    PR1/PR2 10K 25t POTS FEC 935-3704
    R38/R39 IK 1% METAL FILM RESISTORS FEC 946-5170
    R43/R44 47k 1% METAL FILM RESISTORS
    R45 (calculated value) 1% METAL FILM
    R46 4.7 OHM 3W METAL OXIDE RESISTOR FEC 950-4095 (pick a value of about 1/3 of your motor's DC resistance)
    CN3 is wired to the MINI TOGGLE SWITCH on-off-on SPDT FEC 947-3432 ( wiper = centre pin of CN3 and the NO & NC contacts go to the other connections of CN3

    R37; R40; R41; R42 ARE ALL LINKED OUT USING 22 SWG BTC WIRE

    Switching
    A Centre-Off , 2-way mini toggle speed selector switch allows the voltage reference to remain powered up and working to the optimum temperature while in the centre-off position ( otherwise there may be a slight drift for 5-15mins after switch-on) Use a fused, switched mains inlet on your case back panel for a mains-power control.

    Speed choice indication
    R47 is a current limit resistor for the LED speed indicator select for LED’s used. 3K9 was used o the proto build.
    If you want to use LED speed indication then you will need to substitute the Mini Toggle from a SPDT to a DPDT. One pole will be wired to CN3 as mentioned above and the other pole will be wired to CN4. R47 is the ground resistor which is tracked on the PCB to the centre pin of CN4, the other 2 connections on CN4 are also tracked on the PCB and go to +12V and -12V respectively.
    CN4 is connected this way:
    • Centre pin goes to the join of a pair of back to back LED's, the other side of these LED's goes to the wiper of the spare pole on the mini toggle
    • The other 2 pins of CN4 go to normally open and normally closed contacts of the spare pole of the mini toggle, it doesn't matter which way round you wire them.

    How it works: when the speed is selected R47 grounds one of the LED's that is fed from the +12V supply, when the other speed is selected the other LED will light as it is fed from the -12 supply, so it is a kind of flip-flop arrangement, only one LED will remain on at any time.
     
  2. Bemused

    Bemused What's this all about then?

    As ever, thanks
     
  3. sq225917

    sq225917 situation engineer

    I can see my first etching project laid out before me....well almost...

    Thanks Martin
     
  4. neiljadman

    neiljadman Senior Member

    Looks like just what I need to finally get that old RD11E working. Haven't really had a close look yet, but I guess it would be simple enough to add a third preset for 78 +/- 15%?
     
  5. Mark Packer

    Mark Packer pfm Member

    Neil,

    Any idea where could one source a new motor for an RD11S? A good friend of mine has one and it's died. This would mean we could get it up and running again.
     
  6. Mark Packer

    Mark Packer pfm Member

    Also, whilst we're about it...

    Does anyone know which Maxon motors OL sell as their models 100 and 200?
     
  7. 337alant

    337alant Negatively Biased

    Great thread Martin thanks

    Aian
     
  8. sq225917

    sq225917 situation engineer

  9. neiljadman

    neiljadman Senior Member

    Wouldn't any Linn style standard Philips/Airpax/Premotec 9904 motor do?
     
  10. RustyB

    RustyB Registered Ginga

    An LP12 motor is probably the easiest to source, given the upgrading frenzy.

    The RD11 platter can be machined down to correct the gearing.
     
  11. martin clark

    martin clark pinko bodger

    Well it works: here's the lash-up less transformer and output BJT, removed so I can arrange a case for it:

    [​IMG]

    The PCB was courtesy of PulseStudio, a first stab. Having built it there are a couple of very minor tweaks to the layout which would be nice to add, but it does work (!) Further development is going to have to await getting a pulley turned and stuffing the lot into my brother's Rock3 so we can take some performance measurements. Might be a while.
     
  12. starbuck

    starbuck pfm Member

    Looks good - what motor are you using, if you don't mind me asking?
     
  13. 337alant

    337alant Negatively Biased

    Looks very proffessional to me Martin excellent, Ill take one :D

    Alan
     
  14. martin clark

    martin clark pinko bodger

    Starbuck - that's some maxon unit my bro scored (Maxon 2140.93.7). Not necessarily the best thing for the job, just what's to hand. Seems pretty torquey at low speeds though.

    Alan - thanks, but that could be a long time yet, if ever!
     
  15. sq225917

    sq225917 situation engineer

    I'd cover the initial costs for a run of these with mods.
     
  16. shambala

    shambala pfm Member

    Martin,

    I know the feeling, my queue is full. In either case really sweet post. Surely going to work something along these lines at some point. Now back to the wiring.... Hope all is well with all the PFMers out there.

    Cheers
    shambala
     
  17. starbuck

    starbuck pfm Member

    Thanks for the info, Martin. I assume the circuit could be made to work with a Maxon ReMax motor too. I'm not competent enough to build something like this from scratch but, like sq225917, I'd be interested if a run of PCBs is ever done.
     
  18. martin clark

    martin clark pinko bodger

    It should work for any small brusehd motor - just measure the armature dc resistance, pick a sesne resistor value thats less than that and go from there.

    No plans for a pcb run, hadn't even got that far. Paul (pulsestudio) very kindly supplied the example pictured because he was curious and had a margin of space leftover on another project!
     
  19. Chops54

    Chops54 pfm Member

    Paul ( PulseStudio) very kindly put this contoller together for me using a circuit supplied for us by Martin. Paul built the first one for me on vero board and it's been in use ever since on my own turntable which uses the Origin Live DC100 motor. The circuit used was slightly different to the one Martin posted earlier. It's the circuit on the spreadsheet if anyone's interested which is just a simpler version. Anyway Paul then had some printed circuit boards etched and I built up the later, refined circuit. However it didn't work for me so I revised some of the components.


    [​IMG]

    I couldn't get the motor to run slow enough on my Linn Sondek but with these component changes it works fine.It also works equally well on my own turntable with a larger pulley and driving the outside of the platter with no alterations other than adjusting the trim pots.

    I'm not a techie, just a capable person who can use a soldering iron and work stuff out so therefore anyone with half a brain should be able to build a decent controller for their tt.


    I do have a quote somewhere for the cost of producing these boards. I'm not promising anything and I'll need to get Paul's permission but I may be able to organise a group buy if there's enough interest.

    I've just finished another board tonight so I'll post some pics shortly.
     
  20. Fox

    Fox The sound of one hoof clopping

    Tidy layout. Just a suggestion: next version move the two voltage regulators centrally closer together and then the caps can be arranged sideways = lower height profile.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice