Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Vcr head drum speed & phase control.

Status
Not open for further replies.

dr pepper

Well-Known Member
Most Helpful Member
OK, so I have a few salvaged video recorder tape drums.
I'd like to be able to use one in a pov display.
I have vague recollections that the motor in the head is a bldc (this one is), and that it can keep accurate speed and position to track the sync pulses on the tape.
I have one of these on the bench, I sussed out from the controller chip datasheet the powers and grounds, and then with a little trial sussed out which signals are required to make the motor run, but all I can do is get the thing to run flat out.
What is a typical control method for these?, if I ground one of the pins and put a 6kc pwm signal on the other the motor will run and the pwn controls the speed, but you can feel vibration from the motor, reflected by a fluctuating current demand.
 
Sussed it.
I should have fg (frequency gen), pg (pulse gen), start/stop, and speed control voltage connections.
Found out which is speed control voltage and fg (its a pcb track round the motor), start/stop on the chip is permanently wired start internally.
PG which should be 1ppr doesnt appear to be there even though there is 3 diffrent pickups (hall sensors for the bldc driver, pcb track and theres a magnetic pickup on the outside of the motor I thought this would be the 1ppr PG signal, but the 'scope shows nothing), I wonder if this was derived elsewhere in the machine.
FG 20 ppr is as I said a pcb track pickup.
The control voltage varies the speed ok, over a few mins it does change a little, and it gets the chip hot if I go real slow, I wanted 1500 revs which seem doable.
 
The servo needs a phase reference either from Hall Sensors for linear speed control , then a phase detector, loop filter to sync the motor to the POV signal refresh rate so they stay in sync.

Some might use transistor current mirrors to regulate the current which controls torque, others current sense R.


edit I see now you have a 200 ppr signal , now you need a 1 ppr
 
Last edited:
Of the drum there should be a HSWP head switching pulse some where. There is generaly 2 head amps that are alternately switched.
 
I suspect the motor has been divorced from the main board so those signals do not exist.
 
Yes divorce complete.
It has a current sense resistor.
As you say you'd expect a positional indication for phase.
The only o/p I can find is the 20ppr, I just thought however and I'll check this, maybe the pcb track encoder that generates the 20pp signal have 1 or 2 unsymetrical marks to signify head position, so the same 20ppr signal supplies phase info too, I have seen this on another drum which has an optical encoder, there are wider slots in the flag ring to signify head position.
 
There is a 1ppr or maybe 2ppr signal on the connector, it comes right off the sensor and is only 40mv, and its surrounded by a load of noise, probably not worth using, better with a hall effect.

Even though on the freq generatir ring there is a definate 1/2 wide track compared to the rest, but on the 'scope the o/p is perfectly uniform, no shortened section, checked it on my storage 'scope over 50 waveforms.
 
OK then I have precise video head rpm's.
I put the original head motor on one side as I found another in my motors box that has a photo interrupter encoder, and soldering to its collector gives me a 3v squareish wave, which through a schmitt is good and square.
At first I used a 4046 pll with a loop filter, which was then buffered into the analogue voltage control to the head motor, had all kinds of issues stopping it from oscillating, loop filter tuning way out.
Then I remembered doing this a while back with a dc motor and not having a loop filter, the o/p of the phase comparator went straight to a fet driver as it is pretty much pwm, that worked, so I connected the analogue speed control on the motor to V+ (full speed) and then connected the phase comparator o/p direct to the start/stop i/p on the head motor.
Perfect sync at any speed I want up to near the motors max, I just need a freq ref.
Now pic software next.
 
Did you use the PLL on the 1pps or the 20 pps or both? WHich mixer? Type 1, or 2 and did you use 3 ( lock detect)
If using Type 2, speed lock with 1 pps and phase lock with 20 pps for stiffer regulation after Lock detect. But flywheel may have enough enertia so you may not need it.

That 1pps signal may have needed a pullup R.

To generate a pixel clock, Use the same motor PLL but use a divider for the number of columns in your POV plus blanking interval.

Thus you have a sync. harmonic multiplier. for POV pixel column clock to simplify firmware.
If you want smoother characters, use some adjacent pixel averaging like Cleartext.

If you need any really good LED's I have a bunch. As intensity is pulsed with short duty cycles, the average brightness is reduced accordingly. You probably want 10 Cd min. @ 30deg.
 
Last edited:
There are diffrent length pickups for the 20ppr, I have it going to a schmitt trigger after being amped, if you alter the gain of the amp the o/p will pick up on 1pps, 3pps, or the full 20pps, I'm using the latter to speed lock.

I'm using type 2 phase detect, mainly cause the duty cycle of the 20pps is way off 50%, the leading edges however line up, I dont use the phase pulse o/p, pin1.

I tried a pullup and pulldown 10k, killed the o/p altogther, the connection on the bvoard goes direct to the pickup, if its a passive pickup then thats to be expected, I tried holding a 1mH choke up next to the magnet ring, got the same o/p.

Interesting idea about the pixel clock, I'd need a multiplier rather than a divider though as theres only 3 or 4 encoder pulses over the displayed area.

Interesting about your idea of speed locking with 1pps and phase locking with 20pps, were you thinking speed lock with type 2 and phase lock with type 1 in the same chip?, how would I switch with the lock detect and what would I use for a freq ref for the 20pps signal.
 
Last edited:
For RPM stability you need 20 ppr type 2 detector. Using Nx VCO and /N divider, then VCO becomes Pixel Clock.

Since type 1 & 2 are offset 90deg it is hard to combine.

FOr POV orientation stability you need two PLL loops. When Lock detect =1 on 20 ppr occurs , 2nd PLL phase error is added to the loop with Pump Up or Pump Dn using type 2 then float when in sync with first PLL making the main loop correction.

Combine both PLL phase error outputs to regulate both speed and phase using Lock detect to control the gain in one over the other.
 
Yes of course use the vco in the phase lock pll, divide it so I generate a pixel clock.

I need 2 freq ref's then, if my motor rotates at 3000 rpm, 50 revs/sec then I'll need 50hz for the speed loop and 1khz for the phase loop, I could divide this down from one 1khz ref, the counter would need to be synchronous.
 
Last edited:
OK so I got closer.
Instead of 2 loops as per a vcr system I got satisfactory results from one, I added a trimpot and a resistor to the circuit, a fixed dc bias is mixed in from the trimpot to the o/p of the phase comparator integrator to bring the static drive level closer to the target.
Seems to work a lot better now the pll is just trimming rather than generating the control voltage.
 
Thinking about it this is 2 loop, doh.
One of the loops is the trimpot, that sets the speed, and the other is the pll, that keeps phase.
 
The pot is just an open loop set point, which could also be achieved inside the loop by limiting the range of the error loop filter output controlled by the frequency lock detect state.
Lockup time should be <1 sec.
 
Takes more than 1 sec to get to speed, no what you mean though.
 
With a well defined loop, one can use full power to accelerate to desired speed and adjust from velocity error to position error just before arrival of desired speed using some design for velocity feedback and position feedback or frequency and phase using PID where derivative of position is velocity or frequency or speed of the motor.

This is why we used PID loops. Complex non-linear PID feedbacks employ limits to acceleration due to current limts. Where you have simple current limit with pot to control voltage that limits current, a better design goes to max voltage and current just before desired speed to get better lock time. Obviously this cannot be done easily by trial and error, so careful Control System Theory gives optimal results for speed ramp then decay to position error control mode.

Ideal 3 loops are used in 1 integrated loop. Acceleration , velocity and position or phase of POV. Sometimes multiple detector types are used to gain advantages for phase noise during speed lock mode for stability with stiff servo. Then you don't need huge flywheel , which is necessary for low video phase noise but not for POV which is more tolerant.

In the end, one may make a smaller POV unit out of CD motor drive rather than VCR drum drive since mass is low with LEDs but then balance is more critical to avoid wobble from dynamic imbalance. But ideal, is you build it with balanced mass even if LED's is only on one leg.

I like the POV on a bike which requires an index signal and similarily you generate pixel clock from PLL multiplier . But peddle power controls the speed.
130912133247-spoke-pov-bike-wheel-lights-horizontal-gallery.jpg
 
Last edited:
For even higher LED Pixel resolution spin at 7200 RPM or 120 Hz for no flicker with two SMT strips
at 180 deg offset by pitch and interleave strips like interlaced TV and get twice the vertical resolution of Pixels.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top