Thread: VSD power tests

Originally Posted by azzrock

Nice one. But is it possible to get the torque reading from measuring the pressure or flow from the pump instead of from reading the scales.
i was just thinking a flex coupling straight from motor to pump.

bob love your set up. that tacho reads from way back there. luve your lab jack where can i get one.
aaron

aaron, with the method that i mentioned the reading on the scales would indicate foot pounds of torque directly, no calculations necessary, with measuring pressure and flow you would have to know pipe friction, oil temp, viscosity and who knows what else?

shed

Originally Posted by shedhappens

aaron, with the method that i mentioned the reading on the scales would indicate foot pounds of torque directly, no calculations necessary, with measuring pressure and flow you would have to know pipe friction, oil temp, viscosity and who knows what else?

shed

ya good point.shed.

Originally Posted by azzrock

Nice one. But is it possible to get the torque reading from measuring the pressure or flow from the pump instead of from reading the scales.
i was just thinking a flex coupling straight from motor to pump.

A Flex coupling is likely to chew up something that could be significant when assessing smaller motors.

My experience with hydro stuff has not been good, I seem to always end up with puddles of hot oil and stained clothes every time I have used one.

I read up a bit about using a hydraulic dyno. They are very useful because they handle a wide range of Torque/HP and are highly sensitive to rapid changes in Torque and HP but it seems to get the HP both the pressure AND the flow are required, plus the pump efficiency needs to be known. If a fixed displacement pump is used the then we can use manufacturers data (hummm . . . . . if they are anything like air flow rates on DCs I would be a bit worried) and this does not take into account if changes have taken place since the unit was made or last calibrated. If variable flow pumps are used then we also need an accurate flow meter and a set of pump efficiency curves to cover the different flows. To get around needing all this some hydro dynos use a torque arm with strain gauges attached which is exactly what is inside those luggage scales. If you already have this stuff it would definitely worth a go and I would like to see it working.

One beauty of the DeProny brake is that it directly uses fundamental quantities (i.e. mass, length and time) all of which can be easily checked with standards. I haven't checked the newer luggage scale but the older one is accurate to 0.01g compared to a set of standard weights I borrowed from the uni. Not that we need that level of accuracy for what is going on here.

bob love your set up. that tacho reads from way back there. luve your lab jack where can i get one.
aaron

The tach is an elcheapo $20? I bought a few years ago and I have used it heaps. I checked it against a fancy one at uni and it was better than 1% at 100 Hz. The most common problem leading to a gross errors is that it sees unwanted reflections. This can be checked by adding 1 then 2 reflectors and the RPM should double, if it doesn't then there is a problem that needs looking at.

Lab jacks are great - we used them all the time in the labs at Uni. I bought this little Al and SS one from ebay. They are not cheap (even from HK) but they are well worth it as they have a wide range of uses.. I used mine the other day to hold a large cupboard door level while I drilled the holes for the hinges.

Here are the results of measuring HP over a range of VFD frequencies.
Clearly the motor does it's best when running at 50Hz

Are you going to plot frequencies a bit higher - say 80-100Hz, Bob? I for one use that range and would like to know what's left in torque up there.
Many of our motors are designed to run well at 60 Hz, so I didn't think that was pushing the envelope....

Cheers,
Joe

Originally Posted by jhovel

Are you going to plot frequencies a bit higher - say 80-100Hz, Bob? I for one use that range and would like to know what's left in torque up there.
Many of our motors are designed to run well at 60 Hz, so I didn't think that was pushing the envelope....

Cheers,
Joe

Just as a further test idea:

what about plotting like Joe says at higher frequencies, but do this twice. The second time with the fan removed from the rear motor shaft. That would tell us how much energy is wasted by a motor fan spinning at twice its design speed. I personally found that at about 120Hz, a common stock motor produces just enough power to keep itself spinning. If the frequency is pushed any higher, the motor will stall. I wonder how much of the losses are really due to the fan. On my mill I use a motor whose spindle fan has been removed and replaced by two separately powered box fans. And that motor remains usable up to almost 140Hz for small diameter milling cutter work.

Originally Posted by jhovel

Are you going to plot frequencies a bit higher - say 80-100Hz, Bob? I for one use that range and would like to know what's left in torque up there.
Many of our motors are designed to run well at 60 Hz, so I didn't think that was pushing the envelope....

Cheers,
Joe

Yes I will do that. The measurements are very quick, literally 30 seconds, but then loading up the pics and transferring the data to Excel is a bit of PITA.

BTW The motor I am testing specifically says 50Hz on it.

Originally Posted by BobL

.....................

BTW The motor I am testing specifically says 50Hz on it.

The motor manufacturers have had to standardised their production. They do not use different components like rotors or stator iron for 50 and 60Hz countries.

Sometimes they will put both ratings for 50 and 60Hz on the nameplate. And if this is the case, it is interesting to see that the same motor is rated for an about 10 to 15% higher power output at 60Hz, than it is at 50Hz. I wonder why this does not show up in your graph. Maybe it is an older motor, from a time when they still made different designs for 50 and 60Hz?

Hi Bob, it looks to me that your data for each given frequency indicates that as the RPM gets lower the HP goes higher, am I reading that correctly?

shed

Originally Posted by cba_melbourne

Just as a further test idea:

I personally found that at about 120Hz, a common stock motor produces just enough power to keep itself spinning. If the frequency is pushed any higher, the motor will stall. I wonder how much of the losses are really due to the fan. On my mill I use a motor whose spindle fan has been removed and replaced by two separately powered box fans. And that motor remains usable up to almost 140Hz for small diameter milling cutter work.

You could prove or disprove that by putting a small sheet of plastic over the air inlet to the fan, this will
cause the fan to slip and significantly reduce the power required to drive it.
If you have a leaf blower, hold it on full revs and push air inlet into your stomach to block the airflow and the blower will rev its guts out.

shed

Originally Posted by shedhappens

Hi Bob, it looks to me that your data for each given frequency indicates that as the RPM gets lower the HP goes higher, am I reading that correctly?

shed

Hi Shed, Not quite... As the load is increased which means more power, the rpm drops... if we could see current readings we would see current rising as the load increased and the rpm dropped.

Which, brings me to the point I wanted to raise, and that is, the curves would make more sense ( to me at least ) if the hp vs rpm data was at a constant current.

Let's choose rated nameplate current. Set the frequency.. increase the load until we are drawing rated nameplate current and measure rpm and torque at that point.

You could then plot a curve showing hp and torque vs frequency.

Ray

Originally Posted by RayG

Hi Shed, Not quite... As the load is increased which means more power, the rpm drops... if we could see current readings we would see current rising as the load increased and the rpm dropped.

Which, brings me to the point I wanted to raise, and that is, the curves would make more sense ( to me at least ) if the hp vs rpm data was at a constant current.

Let's choose rated nameplate current. Set the frequency.. increase the load until we are drawing rated nameplate current and measure rpm and torque at that point.

You could then plot a curve showing hp and torque vs frequency.

Ray

I understand what you mean but it is tricky to apply the hand brake manually to get the motor to draw a specific current.
However, there is a wing nut on the braking mechanism that I should be able to use to do this.
Tomorrow.

I was a bit busy yesterday. Tuesday morning is my weekly session with the RSL Care fellas, then there was a small family crisis to sort out and when I got home I found two new toys had finally arrived, a plasma cutter and a soldering reworking station, and spent the rest of the day mucking about with these.

Late last night I tried using the Wing nut I described in the previous post to control the back torque but it is just as difficult to apply a back torque such that the motor draws the name plate current with the wing nut as it is with the hand brake. Even slowly approaching the name plate current, the current overshoots the value and then backing off undershoots, if the current eventually hits the name plate value it does not sit at that value but drifts within a second or two. Maybe the heating of the drum and brake band is changing things too quickly. I can eventually get the motor steadily running on the name plate current but by then the paint on the drum is smoking' and I would prefer not to have to do this for every frequency step needed to generate the graph.

I then tried a different approach using the currents I already had for the graph in the previous post. I plotted the range of currents [at each frequency] versus measured HP and derived a polynomial function for HP(Amps). Using the function enables a calculated HP to be determined for any current. The graph below is the HP at the name plate current (3.6A). The 45Hz value seems out place so I will check that one again. I did a simple uncertainty estimate of this method and the tolerance for each point on the graph is about +/- 10%

Interesting. Looks like a reasonably useful range is 3:1 for that motor, assuming 0.5HP is the minimum useful power. 5:1 looks difficult to achieve.

So to get the HP over the frequency range, you need to start with a motor at least 2X the HP that you actually require.

A few motors from different manufacturers would be nice to plot. Pity you're so far away, I have quite a collection of 3 phase motors.

I always thought I'd have to go from a 3HP motor to a 7.5HP VFD driven motor if I ever got rid of the Kopp variator and still need a 2:1 intermediate shaft somewhere. That may even be optimistic.

PDW

Originally Posted by BobL

Something is not right with that plot or with the test setup, or with the VFD settings, or with the VFD itself. The motor power at 60Hz should be about 10% higher than at 50Hz, certainly not lower. I believe that even an older motor should remain more or less constant power from 50Hz up to about 70or 80Hz, before slowly dropping down again and reach zero shaft power at about 120Hz.

At 50Hz and below, the graph looks about right to me.

Be aware, that (almost) every VFD is capable of automatically feeding the motor with a 50 to 100% larger power than rated, for a short time of about 60 seconds. This means that you should let stabilize your test setup for maybe 1.5 to 2 minutes, before taking a reading. Otherwise you could be measuring and plotting this short term overpowering capability of the VFD, instead of continuous duty capability of the motor/VFD combination.