refrigerant EEV stepper motor, unipolar vs bipolar driver.

[Lord Nicoll]

Hello peeps.
It seems a lot of the stepper coils used to drive EEV (electronic expansion valves, a metering device used in refrigeration driven by a stepper moter for those who don't know) are 6 lead unipolar coils. I wanted to just use an A4988 for cheapness and simplicity reasons,  but would require using the stepper coil in the bipolar mode. I can't use the 5 lead ones but is there any real advantage or disadvantage to 6 lead unipolar vs 4 lead bipolar? I'm not sure why most of the them are  6 lead, maybe it's how the system calibrates where it is when it starts up by detecting current when it reaches the end of travel? I have no personally worked a lot with EEV's before and decided to jump in with a "fun" arduino project to make my own EEV control system for a niche application. I'm using 24v for the coils and mechanical logic like contactors and sensors to keep it somewhat compatible with industrial stuff since it's a lotta DIN rails and refrigerant parts.

[sarahMCML]

I'm not sure that you can't use a 5 lead stepper as if it were a 4 lead bipolar one, since it's effectively the same as connecting the two centre leads of a 6 lead stepper together. 6 lead steppers can be used as 4 lead ones by ignoring the centre leads.
If you consider the 2 coils as the 4 arms of a bridge, when the ends are being driven, the centre leads should always be at about half supply voltage, and therefore little if any current should flow between the 2 coils.
I intend to try this out myself tomorrow with a 5 lead stepper as an experiment.

[johansen]

i have never tried using a 5 wire stepper (probably because i've never found one) with a 4 wire driver, but i suspect it will work, especially for the very low power requirements needed to drive an EEV.

the 6 wire steppers are used simply because the cost of 2 extra wires is less than the cost of 2 H bridges compared to the cost of 4 ground referenced transistors or mosfets.

far as i know, most refrigeration systems simply drive the EEV for a certain amount of time until its fully closed, then they open it a certain number of steps. no effort is made to measure the power consumption of the motor (which would be very difficult, given the extremely low efficiency of a stepper motor)

i very highly doubt you will lack sufficient voltage to drive a 6 wire EEV valve from 4 wire stepper driver using the full coil in bipolar mode.

if you were to drive half the coil, it would double the current but the turns are cut in half which results in the same flux density.

[H.O]

A 6-wire stepper can be driven by a bipolar driver by leaving the center taps unconnected (isolated) and use the four ends of the two windings as if it was any other 4 wire motor. Alternatively, use half of the windings and leave one end of each unconnected for lower inducance.
A 5-wire stepper can not be used in this way since the center taps are connected internally which will cause "cross conduction" between the coils when driven in bibolar mode.

[sarahMCML]

Well, I tried out the idea that a 5 wire stepper can be driven by a 4 wire bipolar H-Bridge driver, and it works fine.
I used one of those 28BYJ-48 5V miniature ones, since it is the only one I have, and which I modified into a 6 wire version. I ran it in half step mode from an L298N driver, and monitored the current between the 2 now separate centre windings. The cross conduction between the coils was a fraction of a mA, as I expected it would be, and had no effect on the running of the motor or total current consumption when joined together.

[H.O]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

[johansen]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

As long as you have a quadrature waveform to the two coils it should work
 There is no holding torque needed to drive an eev

[langwadt]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

As long as you have a quadrature waveform to the two coils it should work
 There is no holding torque needed to drive an eev

aren't they spring loaded to close when the power is off?

[sarahMCML]

My L298N driver board has been modified with source resistors providing current feedback to a matching L6506 chopper controller chip. I've used this combo with 4 wire motors and it works really well. I also doubt whether it would be happy trying to drive a 5 wire motor as a 4 wire if using greater than half stepping. I'll have to try that tomorrow with a TMC2208 driver!

Edit: I ran it at 7V to allow for the inefficiency of the L298N, and half step mode.

[johansen]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

As long as you have a quadrature waveform to the two coils it should work
 There is no holding torque needed to drive an eev

aren't they spring loaded to close when the power is off?

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

[Siwastaja]

aren't they spring loaded to close when the power is off?

No. You are confusing EEV with reversing valve. EEV is just a stepper, and the first thing the heatpump does after powerup is to run the stepper with large number of steps to force it to a known position, in one end.

Anyway, unipolar stepper motor has center-taps added compared to the bipolar motor, so unipolar motor can be driven with a bipolar driver with center taps unconnected.

[Lord Nicoll]

Yeah doesn't seem to be any major disadvantages to running the unipolar stepper in bipolar mode (in fact the one I'm testing now is unipolar and running in bipolar mode). Developing my own hardware and software for controlling these may seem odd but I have a niche use and want more control than those cheapish superheat controllers off, such as some contactors and display out for preprogrammed temperature targets that the user can select. I guess nothing left now but to cannibalise some code I found online and pay my electrical engineering friend (I'm a mechanical and manufacturing focused fella) to do the complex coding and then I can actually build it go into the testing phase.

[Lord Nicoll]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

[johansen]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

[johansen]

Yeah doesn't seem to be any major disadvantages to running the unipolar stepper in bipolar mode (in fact the one I'm testing now is unipolar and running in bipolar mode). Developing my own hardware and software for controlling these may seem odd but I have a niche use and want more control than those cheapish superheat controllers off, such as some contactors and display out for preprogrammed temperature targets that the user can select. I guess nothing left now but to cannibalise some code I found online and pay my electrical engineering friend (I'm a mechanical and manufacturing pd fella) to do the complex coding and then I can actually build it go into the testing phase.

If you have enough voltage you will nearly double the torque, and at least one commercially available product already does this to free stuck eev steppers.

[Lord Nicoll]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

Oh, I'm not using R134a, refrigerant of choice will be R1270, this unit will be a very low temperature refrigeration unit, from -60°C low load to about -35°C high load, hence the EEV to allow it to adapt to what it has to do, obviously pressures will vary a bit to allow this, but 220 average, 350 psi is likely to be the realistic most it'll get to. I'm not rating this for 40°C condensing temps so I'll be disregarding th eASHRAE ratings and such. In theory someone at a later date may choice to swap in R404a or R507c, these are still the most common refrigerants people use for these units (these are used to cool CPUs and GPUs and therefore the coldest you can get is absolutely needed, the new refrigerants are just not cold enough and nobody uses them, not to mention the fact they exhibit glide like R449a is almost completely unusable due to it, and the efficiency is just not there). The reason for the EEV is response time, a TEV is simply not fast enough and a computer does not provide a constant load for a capillary tube to be viable at 34cc 1100 watts, the compressor could get flooded at idle with liquid propene.

[langwadt]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

Oh, I'm not using R134a, refrigerant of choice will be R1270, this unit will be a very low temperature refrigeration unit, from -60°C low load to about -35°C high load, hence the EEV to allow it to adapt to what it has to do, obviously pressures will vary a bit to allow this, but 220 average, 350 psi is likely to be the realistic most it'll get to. I'm not rating this for 40°C condensing temps so I'll be disregarding th eASHRAE ratings and such. In theory someone at a later date may choice to swap in R404a or R507c, these are still the most common refrigerants people use for these units (these are used to cool CPUs and GPUs and therefore the coldest you can get is absolutely needed, the new refrigerants are just not cold enough and nobody uses them, not to mention the fact they exhibit glide like R449a is almost completely unusable due to it, and the efficiency is just not there). The reason for the EEV is response time, a TEV is simply not fast enough and a computer does not provide a constant load for a capillary tube to be viable at 34cc 1100 watts, the compressor could get flooded at idle with liquid propene.

any big advantage in using a stepper EEV vs. a pwm on-off EEV ?

[Lord Nicoll]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

Oh, I'm not using R134a, refrigerant of choice will be R1270, this unit will be a very low temperature refrigeration unit, from -60°C low load to about -35°C high load, hence the EEV to allow it to adapt to what it has to do, obviously pressures will vary a bit to allow this, but 220 average, 350 psi is likely to be the realistic most it'll get to. I'm not rating this for 40°C condensing temps so I'll be disregarding th eASHRAE ratings and such. In theory someone at a later date may choice to swap in R404a or R507c, these are still the most common refrigerants people use for these units (these are used to cool CPUs and GPUs and therefore the coldest you can get is absolutely needed, the new refrigerants are just not cold enough and nobody uses them, not to mention the fact they exhibit glide like R449a is almost completely unusable due to it, and the efficiency is just not there). The reason for the EEV is response time, a TEV is simply not fast enough and a computer does not provide a constant load for a capillary tube to be viable at 34cc 1100 watts, the compressor could get flooded at idle with liquid propene.

any big advantage in using a stepper EEV vs. a pwm on-off EEV ?

Well I don't know of any servo or normal motor driven EEVs, they use steppers so they can more accurately meter the refrigerant flow, steppers are an easy and cheap way to do that. By referring to a graph for number of steps vs mass flow, and the refrigerants phase diagram as an f(x) it is possible to very accurately calculate the number of steps needed to open the valve to main the desired temperature, which thanks to the EEV, can be user selected from a small range of possible options.

[sarahMCML]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

Well, I finally got around to setting up the 5 wire test with the modified 6 wire motor. After setting up the TMC2208 driver in half step mode and scoping the motor leads, I find that the action of the chopper current limiting function WOULD drive cross currents between the 2 coils. So while a 5 wire can be driven in 4 wire mode 'full step', other modes cannot, at least if using chopper current controllers. You were correct.

[johansen]

I'm prepared to stand corrected on this but... :-)
Did you run it with a "proper" constant current control IC or did you simply supply the bridge with the 5V that your particular motor coils need across them to reach the rated current and sequenced the inputs for half step drive?

I have a strong feeling it will also depend on what decay mode the constant current controller driver uses.
For example, if it uses slow decay mode where both low side switches are turned on during the off-time of the PWM period it "must" cause weird things. I just can not see how it could work properly and reliably when the two windings are interconnected. But again, I'm prepared to stand corrected.

Try it with a microstepping driver like the A4988 that OP wants to use.

Well, I finally got around to setting up the 5 wire test with the modified 6 wire motor. After setting up the TMC2208 driver in half step mode and scoping the motor leads, I find that the action of the chopper current limiting function WOULD drive cross currents between the 2 coils. So while a 5 wire can be driven in 4 wire mode 'full step', other modes cannot, at least if using chopper current controllers. You were correct.

It may not matter.

I am curious if you can measure much difference in torque between a 6 wire unipolar motor driven by a 4 wire stepper controller with the mid points of the coil open vs the same 6wire stepper with the center points shorted together .

[langwadt]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

Oh, I'm not using R134a, refrigerant of choice will be R1270, this unit will be a very low temperature refrigeration unit, from -60°C low load to about -35°C high load, hence the EEV to allow it to adapt to what it has to do, obviously pressures will vary a bit to allow this, but 220 average, 350 psi is likely to be the realistic most it'll get to. I'm not rating this for 40°C condensing temps so I'll be disregarding th eASHRAE ratings and such. In theory someone at a later date may choice to swap in R404a or R507c, these are still the most common refrigerants people use for these units (these are used to cool CPUs and GPUs and therefore the coldest you can get is absolutely needed, the new refrigerants are just not cold enough and nobody uses them, not to mention the fact they exhibit glide like R449a is almost completely unusable due to it, and the efficiency is just not there). The reason for the EEV is response time, a TEV is simply not fast enough and a computer does not provide a constant load for a capillary tube to be viable at 34cc 1100 watts, the compressor could get flooded at idle with liquid propene.

any big advantage in using a stepper EEV vs. a pwm on-off EEV ?

Well I don't know of any servo or normal motor driven EEVs, they use steppers so they can more accurately meter the refrigerant flow, steppers are an easy and cheap way to do that. By referring to a graph for number of steps vs mass flow, and the refrigerants phase diagram as an f(x) it is possible to very accurately calculate the number of steps needed to open the valve to main the desired temperature, which thanks to the EEV, can be user selected from a small range of possible options.

I was thinking solenoid valves like for example AKV10, I've used those, flow is controlled by ON time in a few seconds period   

[Lord Nicoll]

I doubt it.

Its on the order of 256 to 1024 steps to fully open the 1/8" dia cone, and it has to hold back 400 psi pressure difference reliably. Its going to be a simple leadscrew that does not back drive.

In this case it's more like 220 max PSI, anything over 220 psi (15 bar) will be running outside of the 1:15 compression ratio target. Can push it to 1:20 for lower powerload cooling but the compressor would be under increased loads, not ideal.

Far as I know most eev valves were designed in response to efficiency improvements mandated to r32 and r410 heat pumps, and have to be able to handle 600 or 700 psi high side and open the low side under a near vacuum. So finding a 300psi rated r134 only eev may be difficult, if any ever existed.

Oh, I'm not using R134a, refrigerant of choice will be R1270, this unit will be a very low temperature refrigeration unit, from -60°C low load to about -35°C high load, hence the EEV to allow it to adapt to what it has to do, obviously pressures will vary a bit to allow this, but 220 average, 350 psi is likely to be the realistic most it'll get to. I'm not rating this for 40°C condensing temps so I'll be disregarding th eASHRAE ratings and such. In theory someone at a later date may choice to swap in R404a or R507c, these are still the most common refrigerants people use for these units (these are used to cool CPUs and GPUs and therefore the coldest you can get is absolutely needed, the new refrigerants are just not cold enough and nobody uses them, not to mention the fact they exhibit glide like R449a is almost completely unusable due to it, and the efficiency is just not there). The reason for the EEV is response time, a TEV is simply not fast enough and a computer does not provide a constant load for a capillary tube to be viable at 34cc 1100 watts, the compressor could get flooded at idle with liquid propene.

any big advantage in using a stepper EEV vs. a pwm on-off EEV ?

Well I don't know of any servo or normal motor driven EEVs, they use steppers so they can more accurately meter the refrigerant flow, steppers are an easy and cheap way to do that. By referring to a graph for number of steps vs mass flow, and the refrigerants phase diagram as an f(x) it is possible to very accurately calculate the number of steps needed to open the valve to main the desired temperature, which thanks to the EEV, can be user selected from a small range of possible options.

I was thinking solenoid valves like for example AKV10, I've used those, flow is controlled by ON time in a few seconds period

Oh that really wouldn't work, would not have enough pressure drop across the system, VLT refrigerant systems as mentioned have pretty high compression ratios for refrigeration systems up to 1:15, so you'd still need an orifice tube or capillary tube. The evap in question too requires a constant and stready flow of refrigerant otherwise it'll wildly swing in temp and probably crack the thermal interface material.