Using a UPS on a Synchronous Motor

[FDelGreco]

Not sure if this should go in the tower clock area or electrical horology forum, but it has to do with an electric synchronous motor

Background

I care for a tower clock in a church in the inner city of Cleveland OH. It's a Seth Thomas #16 that an unscrupulous clock repair company electrified in the 1940s. They took away many of the mechanical parts and installed an electric motor and gear reduction box. The motor was missing – probably burned out. I took out the clock, cleaned it up, replaced the motor, and reinstalled everything about nine years ago. The motor is a ~1/10th HP, 1800 rpm synchronous motor that is geared way down.

The Problem

The City of Cleveland power supply is not very reliable in this part of town so there are sporadic power failures. When that happens, I have to climb the clock tower to reset the time – and I'm getting tired of that.

I want to install an Uninterruptible Power Supply (UPS) to keep the clock running when the power is down. Here is my concern: The motor is of the synchronous type that relies on a steady and accurate 60 Hz frequency. A UPS, when activated, generates a synthetic sine wave of 60 Hz but I don't know how accurate it is. I can buy a UPS and put it on my oscilloscope to compare it with the line frequency or connect it to a frequency meter – and maybe have to return the UPS and try another model or brand. So my question is – has anyone used a UPS to power a synchronous motor to maintain accurate time keeping? If so, what brand / model UPS did you use?

Thanks.
Frank

 

[Tim Orr]

Howdy, Frank!

This seems to me to be right in Ken Reindel's wheelhouse: kensclockclinic.com
For years, he's been making devices that allow USA users to run European audio equipment on USA current and vice versa. Part of, as he calls it, his "day job."

Ken has told me that recent developments have made our own AC kinda sloppy in its 60 Hz regulation, and he has developed devices to cope with that too.

Best regards!

Tim Orr

 

[glenhead]

To be clear, no, I haven't ever used a UPS for a synchronous motor. If that shoots the rest of the answer, then there ya go.

I just retired from working in a telecommunications lab. We had enormous distributed UPS systems for most of our equipment, but there were areas where the building UPS didn't extend and we had to use "bricks". Telecom gear also needs a good, steady sine wave to maintain synchronization across a network, especially when it's "pushed" to extreme distances or bad connections or whatever. We moved almost exclusively to CyberPower UPSs for our bricks. They're relatively inexpensive and wowsers, do they perform. We never saw any evidence of frequency drift from them. They have a line of bricks that work with Apple products and some other stuff that requires a pure sine wave rather than a simulated one, so if you get one of their bricks with the PFC function (such as the CP1000PFCLCD) it'll fit the bill. A simulated sine wave from a well-made brick has almost no chance of causing issues, but the PFC function takes care of the "almost".

Hope this helps.
Glen

 

[FDelGreco]

Glen: Thanks! Frank

 

[ElectricTime]

Pretty good article on UPS - I think you would want either a line interactive or an offline UPS - finding out the frequency stability when on battery of any of these might be hard, however.

We do make a "Stop" tower clock controller that would reset any AC motor after a power failure - it does take a while - it waits until the time is correct - the p/n is CTRL-99B-STOP - it will also handle daylight savings.

 

[Schatznut]

There are two types of error - random and systematic.

Random errors, over a long period of time, will average to zero. This is what is seen in the electric power grid. There's controversy about maintaining the accuracy of the grid frequency, which traditionally has been excellent, with utilities wanting to reduce their responsibility for that accuracy.

Systematic error is the predominant error in quartz oscillators. This is the tendency of an oscillator to deviate a fixed amount from the nominal frequency - they either want to run fast by a fixed amount or slow by a fixed amount. They don't tend to wander, except due to the effect of ambient temperature changes. Certain cuts of quartz crystals are less sensitive to temperature changes, and for the ultimate in minimum drift, there are temperature-controlled crystal oscillators (TCXOs) or oven-controlled crystal oscillators (OCXOs).

If your UPS frequency source is quartz-based, it will depend on how close to the nominal frequency the oscillator runs, and how stable is the temperature of the ambient environment. I don't know much about UPS devices, but I'd be very surprised if the better ones don't have quite accurate and stable time bases.

 

[FDelGreco]

Thanks. I think I might call the two major UPS makers and have them put their reputations on the line.

Frank

 

[Cheezhead]

Post deleted.

 

[Cheezhead]

Just for fun I connected a 120 volt alarm clock to a cheap 70 watt cigar lighter square wave inverter. The clock is already 12 minutes slow in 7 hours. I will later check out my 150 and 300 watt inverters to see if they hold closer to 60 cycles.

With a similar setup powered from a battery you can avoid some trips up the tower when a power outage has a short time duration. With good luck of the draw you might get a more accurate inverter.

An automatic transfer switch can be set up with 120 vac and 12 vdc relays. I have a rough outline of a circuit to indicate that it is possible. It's vital that the transfer switch must reliably isolate either 120 volt source from the other during a transfer both ways with loss and return of line power. Fuse everything! With a relay having NO and NC contacts the term used is non-overlap but much more non-overlap time is most likely needed than what a single relay will provide. A relay may or may not have overlap.

 

[FDelGreco]

If a UPS ran the clock a little faster rather than slower when the power was out, if the clock was fast after a power outage, I could just disconnect it from the UPS until the world caught up with the clock, thereby eliminating a trip up the tower.

Frank

 

[Wimberleytech]

A "standby UPS" only activates the inverter when power is lost. In the mean time, it uses the mains power to drive the load and thus you benefit from the frequency of the mains (what you want). Most (if not all) inexpensive UPS are standby UPS, so buy one from APC and you should be good to go.

 

[FDelGreco]

Wim:

I'm aware of all that.

Frank

 

[Wimberleytech]

Wim:

I'm aware of all that.

Frank

Good. Let me answer in another way. Any UPS inverter output frequency is likely to be based on a crystal, so the accuracy will be good enough for those periods when the power is interrupted (unless it is days and weeks). Go buy an APC inverter--from Amazon--and test it. My bet...it will be fine.

BTW, I have UPSs all over my house (TV, computers...). I doubt that I am unique. If you have one, go measure it.

 

[FDelGreco]

I have three UPSs around my house. One on each PC and one on the TV/cable box so I can finish watching a movie when the power goes out!

Frank

 

[Cheezhead]

My 150W inverter is already running the clock slow, 12 min slow in about 2 hours. Will try the 300 watt inverter later; it's located eleswhere.

I wonder if someone well versed in electronics could speed it up.

 

[Wimberleytech]

My 150W inverter is already running the clock slow, 12 min slow in about 2 hours. Will try the 300 watt inverter later; it's located eleswhere.

I wonder if someone well versed in electronics could speed it up.

Can you post some details of your inverter? It is surprising to me that it is running that slow.

 

[Cheezhead]

It's a 140, not 150 watt inverter. The label says PC140 Power Inverter with no company name. It's got to be over 25 years old and is powered from a battery charger with output voltage measured as 13.0 volts. I suppose the charger is half wave rectified but it does have a small filter cap that I added some years ago. If 12 min slow in two hours, that is 10% slow to indicate 54 cycles. When actuating the alarm buzzer, the tone is about one note lower from the setup as compared to the tone at 60 cycles from the wall outlet.

I have no portable pure DC supply; would need to bring the apparatus to my car for that. I will do that.

 

[Wimberleytech]

It's a 140, not 150 watt inverter. The label says PC140 Power Inverter with no company name. It's got to be over 25 years old and is powered from a battery charger with output voltage measured as 13.0 volts. I suppose the charger is half wave rectified but it does have a small filter cap that I added some years ago. If 12 min slow in two hours, that is 10% slow to indicate 54 cycles. When actuating the alarm buzzer, the tone is about one note lower from the setup as compared to the tone at 60 cycles from the wall outlet.

I have no portable pure DC supply; would need to bring the apparatus to my car for that. I will do that.


View attachment 719149 View attachment 719149

No surprise at the inaccuracy...this is ancient technology...lol

 

[Cheezhead]

A two hour somewhat crudely conducted test done twice showed that the 140 watt 12 vdc to 120 vac inverter frequency increases as input voltage is increased. My 10 amp auto battery charger has a low (12.6 volt) and high (14.1 volt) setting with readings made after the battery voltage settled. Set at low the clock was 6 minutes slow in two hours. Set at high the clock was 4-1/2 minutes slow in two hours.

Effect on inverter output frequency with an added electrical load in parallel with the clock is unknown. That requires another test but I will let the OP take it from here if he will.

 

[ElectricTime]

My 150W inverter is already running the clock slow, 12 min slow in about 2 hours. Will try the 300 watt inverter later; it's located eleswhere.

I wonder if someone well versed in electronics could speed it up.

this is the problem henry warren had when he hooked up synchronous clocks to the Boston power grid. I would think more likely than not your inverter depends on an RC circuit for frequency stability - which are poor for stability but fine for most applications.