http://search.digikey.com/scripts/DkSea ... 520NPBF-ND
First, you may recognize that I simplified the design a bit by using a Full H-Bridge L298 IC.
The Full bridge lets me run single phase pumps that aren't center tapped, as well as get more torque out of those that are. Both circuits in the L298 are used in parallel to increase the power handling capability.
Instead of using Nand gates to power on and off the control signals, I am using the built in enable pin on the bridge. Since the bridge is turned on or off directly I was able to dispense with the Nand (NOT) gates near the end of the circuit. I could have done away with the Nand gates entirely, however I chose to keep the first two and connect them to an interlock circuit. With this particular H-Bridge IC, having both inputs set to the same value shorts the motor outputs to each other, creating a braking effect. I have taken advantage of this to create an emergency stop switch.
Other improvements include TVS diodes near the motor to protect the circuitry.
- (79.39 KiB) Downloaded 4687 times
I finally finished a pumping station based on Alexi’s controller and a TPH-055 purchased off of EBay. All of the vacuum fittings are from Lesker’s; the main gauge in the system is an MKS-901P which bottoms out at 10-5 mTorr (also another snag off of EBay). The gauge bottoms out about five minutes after turning on power to the turbo pump.
Most of the parts for the controller were purchased from Digikey and Radio Shack. I also included a cheap frequency counter on the front panel driven off of pin 3 of the 555 timer.
The only thing that I did to the pump was to use some fine grain sandpaper (as Carl suggested in his post) to smooth out some of the dings in the flange of the pump. I actually purchased two pumps of off EBay and both of them are fully operational after following the cleaning instructions (to get rid of the black gunk in the pump) and smooth out the divots in the top vacuum flange.
I will probably replace the grease reservoir in the pump too since it is not too expensive and seems like a modest investment to keep the pump in good working order.
- (152.66 KiB) Downloaded 4687 times
The proper connectors for the TPH-055 pump and TCP 035 and TCP 015 controllers are:
Pump body connector: 12-8p or equivalent
Pump side of cable: 12-8s or equivalent (PT06E-12-8S)
http://www.alliedelec.com/search/produc ... U=70143407
Controller side of cable: Harting 09-06-231-6822
http://www.mouser.com/ProductDetail/HAR ... dkYPFT0%3d
Controller board connector: Harting 09-06-131-6922
http://www.mouser.com/ProductDetail/HAR ... RPspXXI%3d
Harting connector data sheet:
Good luck with the pump controllers, I'll try to build some boards with my CNC mill as time permits.
http://www.ebay.com/itm/300441579297?ss ... 1497.l2649
turbo-control.brd - eagle cad board design
turbo-controlbd.nc - gcode drill file for holes units in mm
turbo-controlbot.nc - gcode mill file for traces units in mm
Pics as following:
3x PCB after milling
hall sensor output
3x assembled board
WARNING _ WARNING: Do not close the top jumper above the turbo connector, everything will catch fire (to be fixed in rev 2) other then that, everything works fine although it only reached 1.3kHz rotation speed.
- (338.13 KiB) Downloaded 4687 times
- (289.21 KiB) Downloaded 4687 times
- (281.43 KiB) Downloaded 4687 times
- (322.68 KiB) Downloaded 4687 times
- (442.64 KiB) Downloaded 4687 times
- (299.16 KiB) Downloaded 4687 times
- (391.72 KiB) Downloaded 4687 times
- (309.52 KiB) Downloaded 4687 times
- (355.99 KiB) Downloaded 4687 times
- (423.63 KiB) Downloaded 4687 times
[The extension nc has been deactivated and can no longer be displayed.]
[The extension nc has been deactivated and can no longer be displayed.]
[The extension brd has been deactivated and can no longer be displayed.]
Fusion is the energy of the future....and it always will be
Retired now...Doing only what I want and not what I should...every day is a saturday.
I just want to point out a few things which are wrong with Chris Trent's and yours controller circuit.
1. The hall sensor output (pin G) needs a pull-up resistor about 10k to +5V. Without this resistor the hall sensor signal looks horrible and might not trigger the 555 timer reliably.
2. Driving the turbo pump motor which has a sinusoidal back-emf proportional to its frequency with fixed amplitude square wave from an H-bridge is like trying to fit a square peg into a round hole. Something got to get hot: either the H-bridge or the pump's motor. You don't want either to get hot. Simple solution is to connect a 5 to 10 ohm resistor in series with the motor or in series with the H-bridge power.
3. If you are using the whole winding (pins B and C) then you need to double the voltage (at least 50V) in order to get full speed. Alternatively, you can use a halve-winding (pins A and B or A and C) with a 30V power supply.
4. The speed control circuit will work way better if you separate wires for the hall sensor from the wires for the motor windings (use two cables, not one) and use shielded cables.
5. Don't try to stop the pump by shorting the winding as Chris suggested. There is enough energy stored in the rotor spinning at full speed to melt the windings.
I hope that all those suggestions will help you to improve your controller.
Some feedback from Rich Normand:
I tried both Alexis and Chris circuits and settled on Alexis but with
replacing the current limiting resistor with two 12V H4 halogen lamps and
NTRs all in series. At the start the NTRs have large resistance and the
lamps low resistance, limiting the strat up current surge. As the pump revs
up the NTRs warms up and get down to about 2 ohms while the lamps start to
glow brightly thus limiting current and changing thier resistance depending
on current. At top speed they go back cold to a few ohms and the running
current keeps the NTRs slightly warm. Best of both worlds.
I also built Chris circuit and could not get my pump to go to top speed. I
think the issue is although the L298 chip can reverse the applied voltage
polarity the 28V for 100% duty cycle, but is only applied to the two coils
in series by the out2, out 3 and out1 out4 sequence, instead of in Alexis
circuit where the full 28V is applied to each individual coils in sequence.
Am I wrong here? If so running at twice the voltage might work but I did not
want to try!!!!
If the previous is true, grounding the center tap will result in mahem as it
will be at half voltage at any given time while the virtual ground is being
flipped at each half rotation.....
Perhaps a modification to take advantage of both topologies would be the
answer? Alexi's waste half the duty cycle by powering one coil at the time
so a merging of both might work even better?
Did not want to post any of this till checking with you since you have done
quite a bit of work with these.
I found the problem that was causing the driver to catch fire. I am using a 1ohm current limiting resistor, and each pump coil has a resistance of about 1.5 ohm. With both bridge drivers in parallel the maximum continuous current permissible through the bridge is 4A. At 24V with the center point grounded, if power is applied when the pump is at standstill, the bridge will be carrying 10A, which is why it will catch fire. If you ground the center tap and start the driver at 5V, it starts up just fine. For the long term we should have some manner of current limiting device in series with the motor power.
Obviously the motor needs a constant V/Hz drive scheme to limit start up current, which this simple driver is not capable of providing without a 10 ohm or so series resistor that will still dissipate power when the pump reached full speed.
What do you all think about adding a LM317 linear regulator in a current limiting configuration in series with the 24V bus? With a 1ohm feedback resistor this will limit current to 1.25A at all times and should have minimal losses at higher speeds.