universal uninterruptible power supply

What would it take in the design of a UPS to make it work correctly and safely under conditions where the incoming power could have either wire grounded or
neither wire grounded (but there will always be a separate equipment grounding conductor), and will operate over the nominal voltage range of 200 to 240 with either 50 or 60 Hz input?
A big question might be what kind of output power it would have, such as if either wire can be grounded, a specific wire would be grounded, or neither wire would be grounded. Would the output always be a specific design, or could it follow what the input is. Would it be able to safely have a bypass mode, or would that have to be omitted?
Would this require input and/or output isolation transformers, or could it be done without transformers?
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

A fixed frequency design for either 50 or 60 Hz would be pretty standard. I haven't seen many (any?) UPS with a frequency switch or auto frequency setting - they all tend to be for a single fixed frequency.

My first thought would be for either a ferro-resonant or a double conversion "online" type - without any bypass mode.

Of course it could be done without transformers in a number of ways - an mg set springs to mind. And the transformers themselves, if used, need not be at mains frequency. But why would you want to avoid transformers?
-- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| snipped-for-privacy@ipal.net wrote: |> What would it take in the design of a UPS to make it work correctly and safely |> under conditions where the incoming power could have either wire grounded or |> neither wire grounded (but there will always be a separate equipment grounding |> conductor), and will operate over the nominal voltage range of 200 to 240 with |> either 50 or 60 Hz input? | | A fixed frequency design for either 50 or 60 Hz would be pretty | standard. I haven't seen many (any?) UPS with a frequency switch or auto | frequency setting - they all tend to be for a single fixed frequency.
A great many I have seen do have 50/60 Hz frequency agility. There are countries with 220-240 volt (L-N) systems at 60 Hz. Brazil is one of them (for the portion of the country running on 60 Hz).
However, I would consider it more interesting to see how the engineers deal with the grounding aspect.
|> A big question might be what kind of output power it would have, such as if |> either wire can be grounded, a specific wire would be grounded, or neither |> wire would be grounded. Would the output always be a specific design, or |> could it follow what the input is. Would it be able to safely have a bypass |> mode, or would that have to be omitted? | | My first thought would be for either a ferro-resonant or a double | conversion "online" type - without any bypass mode.
So just eliminate the bypass mode, and provide some kind of switch to select whether the ground is connected between the two 100-120 volt output inverters (for a system like in the USA) or on one end (for a system like in UK).
I'll assume "double conversion online" type.
| Of course it could be done without transformers in a number of ways - an | mg set springs to mind. And the transformers themselves, if used, need | not be at mains frequency. But why would you want to avoid transformers?
Why spend the extra cost and suffer the electrical waste if it's possible to do without them (either the transformer or the motor-generator set)?
I would envision a UPS with 2 inverters in the 100-120 agile voltage range which could be connected in series, with the ground switchable between the middle between them, or on one end. With a bit more in the switching, the 2 inverters could even be wired in parallel to get 100-120 volts out instead of 200-240. No transformer needed. No mg set needed.
How complex would that switching have to be? What would be involved in making it safe enough to get it listed by various listing entities (UL, TUV, etc). Would latching relays be good enough?
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

Thanks for that - only goes to show, I have never seen one with frequency agility, the ones I have seen have either been 50 or 60 Hz. Of course it would be easy enough to sample the incoming supply, on connection, store that value and use it to set the inverter output frequency. Plus, for a stepped output unit (not a "true" sine wave output unit) the frequency change within the inverter wouldn't have a great design implication.

A transformer operating at a frequency rather higher than mains is pretty cheap, small and efficient and, IME, very reliable..

But I would suggest a lot more expensive and larger and less reliable than one inverter plus 20+KHz transformer.

The switching itself would be pretty trivial - a little more complex would be the sensing to ensure fail-safe. Hence mostly down to relay design - eg extra contact sets that prevent the inverter(s) operating (or operating at significant power levels) if the required safe operating conditions hadn't been achieved during the initial phase of transfer. -- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| A transformer operating at a frequency rather higher than mains is | pretty cheap, small and efficient and, IME, very reliable..
Of course a higher frequency means a smaller transformer. But then, you don't have 50/60 Hz output from that transformer.
|> I would envision a UPS with 2 inverters in the 100-120 agile voltage range |> which could be connected in series, with the ground switchable between the |> middle between them, or on one end. With a bit more in the switching, the |> 2 inverters could even be wired in parallel to get 100-120 volts out instead |> of 200-240. No transformer needed. No mg set needed. | | But I would suggest a lot more expensive and larger and less reliable | than one inverter plus 20+KHz transformer.
And I would suggest 20 kHz would have issues providing power for other devices, including computer PSUs.
If output needs to be 50 and/or 60 Hz, and sine wave, now what?
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

For a 20kHz transformer, its output is going to be converted to (typically several hundred volts) DC and then fed to an output stage that generates the 50/60Hz required.

Design the output stage to match the requirement. A true sine wave 50/60Hz oscillator linked to a power amplifier supplied from the DC obtained from the 20kHz transformer will do nicely, if sine wave output is required. Although, often, stepped "sine wave" output is going to be cheaper, more efficient and acceptable.
-- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| snipped-for-privacy@ipal.net wrote:
|> |> | A transformer operating at a frequency rather higher than mains is |> | pretty cheap, small and efficient and, IME, very reliable.. |> |> Of course a higher frequency means a smaller transformer. But then, you |> don't have 50/60 Hz output from that transformer. | | For a 20kHz transformer, its output is going to be converted to | (typically several hundred volts) DC and then fed to an output stage | that generates the 50/60Hz required.
So how many total stages are involved with all this, and how much power loss is involved as a result? You have to convert the power to 20 kHz first just to use that lightweight transformer. Then you have to convert it back to ultimately 50/60 Hz again. It would seem to me that all that conversion would make the reduced transformer loss meaningless.
|> |> I would envision a UPS with 2 inverters in the 100-120 agile voltage range |> |> which could be connected in series, with the ground switchable between the |> |> middle between them, or on one end. With a bit more in the switching, the |> |> 2 inverters could even be wired in parallel to get 100-120 volts out instead |> |> of 200-240. No transformer needed. No mg set needed. |> | |> | But I would suggest a lot more expensive and larger and less reliable |> | than one inverter plus 20+KHz transformer. |> |> And I would suggest 20 kHz would have issues providing power for other |> devices, including computer PSUs. |> |> If output needs to be 50 and/or 60 Hz, and sine wave, now what? |> | | Design the output stage to match the requirement. A true sine wave | 50/60Hz oscillator linked to a power amplifier supplied from the DC | obtained from the 20kHz transformer will do nicely, if sine wave output | is required. Although, often, stepped "sine wave" output is going to be | cheaper, more efficient and acceptable.
Stepped sine wave may well work fine, if there are enough steps involved. But supply power is 220 to 240 V (200 V in Japan), at 50 or 60 Hz, and the output needs to be the same.
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

This is a UPS that we are discussing - at some point the input has to be converted to DC and the output stages powered by DC. As the proliferation of SMPSU shows, the 50/60Hz transformer has near enough disappeared for the former. The same design arguments apply to the latter = there are considerable advantages in stepping the battery voltage up to a high enough DC voltage, using a 20kHZ inverter (with 20kHz transformer) and follow it with a transformerless 50/60Hz output stage - than have a 50/60Hz inverter (with 50/60 HZ transformer).
-- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| snipped-for-privacy@ipal.net wrote:
|> |> |> |> | A transformer operating at a frequency rather higher than mains is |> |> | pretty cheap, small and efficient and, IME, very reliable.. |> |> |> |> Of course a higher frequency means a smaller transformer. But then, you |> |> don't have 50/60 Hz output from that transformer. |> | |> | For a 20kHz transformer, its output is going to be converted to |> | (typically several hundred volts) DC and then fed to an output stage |> | that generates the 50/60Hz required. |> |> So how many total stages are involved with all this, and how much power loss |> is involved as a result? You have to convert the power to 20 kHz first just |> to use that lightweight transformer. Then you have to convert it back to |> ultimately 50/60 Hz again. It would seem to me that all that conversion would |> make the reduced transformer loss meaningless. | | This is a UPS that we are discussing - at some point the input has to be | converted to DC and the output stages powered by DC. As the | proliferation of SMPSU shows, the 50/60Hz transformer has near enough | disappeared for the former. The same design arguments apply to the | latter = there are considerable advantages in stepping the battery | voltage up to a high enough DC voltage, using a 20kHZ inverter (with | 20kHz transformer) and follow it with a transformerless 50/60Hz output | stage - than have a 50/60Hz inverter (with 50/60 HZ transformer).
So: AC -> DC -> 20 kHz AC -> step-up -> high V 20 kHz AC -> high V DC -> 50/60 Hz AC Is that right?
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
** FUCK OFF
YOU ASININE RADIO HAM IMBECILE !!!
..... Phi l
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote: <Snip>

Almost. I'd suggest:
The first bit would be-
AC -> DC -> 20 kHz AC -> step-DOWN -> Very LOW voltage 20 kHz AC -> Very LOW V DC ->
This produces the very low voltage DC needed to (1) charge/maintain the battery voltage and (2) provide power for the output inverter when mains is available.
All the above is, is a SMPSU, not very different to the one in any desktop computer.
The second bit would be-
Very LOW V DC -> 20 kHz AC -> step-up -> "high" V 20 kHz AC -> "high" V DC -> 50/60 HZ AC
Which is just what you would find in most boats/RVs etc, to provide a mains supply from one or more deep-discharge batteries.
-- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
| snipped-for-privacy@ipal.net wrote: | <Snip> |> So: AC -> DC -> 20 kHz AC -> step-up -> high V 20 kHz AC -> high V DC -> |> 50/60 Hz AC |> Is that right? |> | | Almost. I'd suggest: | | The first bit would be- | | AC -> DC -> 20 kHz AC -> step-DOWN -> Very LOW voltage 20 kHz AC -> Very | LOW V DC ->
How well filtered would the first AC -> DC part need to be? I would think it might not matter, much.
| This produces the very low voltage DC needed to (1) charge/maintain the | battery voltage and (2) provide power for the output inverter when mains | is available. | | All the above is, is a SMPSU, not very different to the one in any | desktop computer. | | The second bit would be- | | Very LOW V DC -> 20 kHz AC -> step-up -> "high" V 20 kHz AC -> "high" V | DC -> 50/60 HZ AC | | | Which is just what you would find in most boats/RVs etc, to provide a | mains supply from one or more deep-discharge batteries.
Ultimately this can be encapsulated:
AC at supply voltage/frequency -> [first bit] -> DC at battery voltage Appropriate parallel/switched operation with batter DC at battery voltage -> [second bit] -> AC at utlization voltage/frequency
Now, back to my original issue, with these "encapsulated" modules doing the AC (supply) -> DC (battery) -> AC (utilization)
My concern is the practicality of a universal UPS that works on an AC power supply in the 50..60 Hz range, with any one of these systems:
1. Two wire 200..240 V grounded at wire A. 2. Two wire 200..240 V grounded half way between wire A and wire B 3. Two wire 200..240 V grounded at wire B.
With the _possibility_ that it can also at least supply output power like:
4. Two wire 100..120 V grounded at wire A. 5. Two wire 100..120 V grounded at wire B.
The design with 20 kHz AC step-down and step-up stages doesn't seem to apply to dealing with the utiliztion voltage systems here, EXCEPT that it could be a semi-split design, where the inverter is a single unit up to the point where the final "high" V DC is converted to the final 50/60 Hz AC. If the latter is split into two parts, along with that final DC voltage being chosen for this, it could still produce TWO isolated 100..120 V AC outputs that could be used to produce system #2 above, as well as #4 and #5.
There are other issues. For example if the UPS has domestic outlets, there will be an issue of which conductor is allowed to be grounded. With outlets of the style used in Argentina, Britain, France, India, and USA (for 120V, not for 240V), a specific conductor must be the grounded one (let's call it wire A). If the UPS were to output any other system of power, it would be operating in an unsafe way with respect to the expectations in that country.
However, if a correctly wired AC supply were given, it could "learn" what the correct way is and output exactly the same. I don't know if that would enough for the appropriate equipment safety listings.
Consider the case of Argentina and Australia. They use the same outlet/plug design, but have the current carrying conductor that is to be grounded in the opposite configuration. Consider a UPS constructed with this kind of plug and outlet. It could learn which conductor is grounded and do the same for its outlet. That's case #1 and #3 above. From an engineering perspective, how hard would it be to make it safe under the conditions that someone does in fact, actually plug it in to a live outlet in the host country (as opposed to plugging it in one, then taking it to the other and operating it standalone using its charged battery power, or plugging it into an incorrectly wired outlet).
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

The main question is whether the savings in cost resulting from only having to make (and support) one product would be greater than the increased cost of a more complex design - and getting that one design through all the approvals process of all the countries concerned. Not forgetting the costs resulting from the delay in getting such a product to market as a result of all of these approval processes.
As we appear to be about to enter into a protectionist spiral - now may not be a good time to invest in "universal" designs...
-- Sue
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:
<snip>

With a transformer isolating the input side from the output side, all questions of grounding this side or that side, become non-issues. Transformer isolation can allow the output circuitry to be designed completely floating, allowing you to ground or not ground either leg without regard to incoming.

Running two inverters in series or parallel is almost *always* harder to do than running a single, larger one. Getting the triggering and output stages to 'play nice' and always work together reliably is harder then just getting larger power stage components.
daestrom
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote: | snipped-for-privacy@ipal.net wrote:
| <snip> |>> Of course it could be done without transformers in a number of ways |>> - an |>> mg set springs to mind. And the transformers themselves, if used, |>> need |>> not be at mains frequency. But why would you want to avoid |>> transformers? |> |> Why spend the extra cost and suffer the electrical waste if it's |> possible |> to do without them (either the transformer or the motor-generator |> set)? |> | | With a transformer isolating the input side from the output side, all | questions of grounding this side or that side, become non-issues. | Transformer isolation can allow the output circuitry to be designed | completely floating, allowing you to ground or not ground either leg without | regard to incoming.
True. But at 50/60 Hz, this is a big transformer. And I do not see a need to produce a system type different than what is coming in. If the derived system is made to match the supply system, then engaging the bypass would not change the nature of the system.
|> I would envision a UPS with 2 inverters in the 100-120 agile voltage |> range which could be connected in series, with the ground switchable |> between the middle between them, or on one end. With a bit more in |> the switching, the 2 inverters could even be wired in parallel to get |> 100-120 volts out instead of 200-240. No transformer needed. No mg |> set needed. |> | | Running two inverters in series or parallel is almost *always* harder to do | than running a single, larger one. Getting the triggering and output stages | to 'play nice' and always work together reliably is harder then just getting | larger power stage components.
So how can you produce a "grounded center tap" system with a "single, larger" inverter?
If the derived system is _not_ a "grounded center tap" type system, then that would complicate the bypass switching, because that would mean a sudden shift in grounding relations when the bypass is engaged.
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

I never said the transformer had to be 50/60. Take the line, rectify it, then invert it at high frequency and drive the primary of a *small* transformer. Take the secondary, rectify it and drive a power inverter at 50/60. Complete isolation and no 'big transformer'.

Ah, for that you *would* use a large 50/60 transformer as the output stage. Drive at 240 and have center-tapped secondary. At least that's one way. There may be others.
So that would be a compromise of either a big bulky 50/60 transformer or more-difficult/ less-reliable inverter design.

If the service is for a grounded center tap, then simply ground the center of the inverter output (either center-tap of power 50/60 transformer or leg between your two-inverter design). With *just* a ground connection on the secondary side, there isn't any interaction with the mains.
Obviously if the unit is to be used in a variety of service, the exact grounding of the output has to be user configurable. But if the output is isolated from input using either a high-frequency or low-frequency transformer, you can ground any *one* point of the output with no ill-effects.
(well, grounding one side of a power inverter so it 'looks' like a grounded-on-one-side UK supply, *may* cause some noise issues and EMI that affects other equipment, but the basic inverter would still be operational).
daestrom
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:
| I never said the transformer had to be 50/60. Take the line, rectify it, | then invert it at high frequency and drive the primary of a *small* | transformer. Take the secondary, rectify it and drive a power inverter at | 50/60. Complete isolation and no 'big transformer'.
What is the efficiency of all this conversion to and from a higher frequency?
|> So how can you produce a "grounded center tap" system with a "single, |> larger" inverter? | | Ah, for that you *would* use a large 50/60 transformer as the output stage. | Drive at 240 and have center-tapped secondary. At least that's one way. | There may be others.
An "other" way I have in mind is a pair of synchronized 100-120 V inverters wired in series. Ground would be connected to the center when/where that kind of system is needed or desired.
| So that would be a compromise of either a big bulky 50/60 transformer or | more-difficult/ less-reliable inverter design.
Are 100-120 V inverters less reliable? For a given UPS capacity this would involve 2 inverters at half the capacity, of course.
|> If the derived system is _not_ a "grounded center tap" type system, |> then that |> would complicate the bypass switching, because that would mean a |> sudden shift |> in grounding relations when the bypass is engaged. | | If the service is for a grounded center tap, then simply ground the center | of the inverter output (either center-tap of power 50/60 transformer or leg | between your two-inverter design). With *just* a ground connection on the | secondary side, there isn't any interaction with the mains. | | Obviously if the unit is to be used in a variety of service, the exact | grounding of the output has to be user configurable. But if the output is | isolated from input using either a high-frequency or low-frequency | transformer, you can ground any *one* point of the output with no | ill-effects. | | (well, grounding one side of a power inverter so it 'looks' like a | grounded-on-one-side UK supply, *may* cause some noise issues and EMI that | affects other equipment, but the basic inverter would still be operational).
The goal I'm looking at is a UPS that can be connected to either kind of power system. In the case of single ended, either end might be grounded (so it can be used in continental Europe and similar places with reversable plugs). In the case of center tap grounded, only the 2 hot wires are coming in, and the neutral is not (pure 240V American style). The output should match the input system, and it should be automatically done correctly and safely. That means the grounding conductor would be connected at the appropriate point on the output. In the case of American center tapped 240V, the point between the two inverters would be grounded, but no neutral grounded conductor would be needed since the intent is to provide 240V (200V in Japan) the same way it is available from the supply. In the case of single ended, the correct output conductor has to be grounded such that when the bypass switch is engaged, the grounding status does not change. There should be NO "ground shift". That would mean opening the grounding conductor and closing the neutral conductor at the same time. But should this be an open transition to ensure there is no issue with "downstream double bonding"?
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
snipped-for-privacy@ipal.net wrote:

<snip>
No, 120V inverters are not inherently less reliable. But two of any thing instead of one means twice as many components to fail. Adding a second inverter probably reduces the overall products reliability by more than adding a transformer to a single inverter.

:-/, *automatically* reconfiguraing the ground sounds like trouble. That implies some switching/relaying in the ground path and I doubt that would fly past codes.
And how can the unit sense what the input configuration is? Unless you supply it with a separate ground for sensing what (if any) input phase is grounded, I don't see how it can tell.
It would probably be easier to leave the grounding to a 'user configurable' setting. Have the user configure the output inverter ground connection before installing the unit and leave it connected all the time. Then when you switch from bypass to UPS output, there is no ground-shifting either.
daestrom
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
wrote:
| :-/, *automatically* reconfiguraing the ground sounds like trouble. That | implies some switching/relaying in the ground path and I doubt that would | fly past codes.
Listing codes, like UL?
| And how can the unit sense what the input configuration is? Unless you | supply it with a separate ground for sensing what (if any) input phase is | grounded, I don't see how it can tell.
There would be an incoming EGC. It should detect how that EGC relates to the TWO wires of input. If it fits on or nearly on a point between the 2 input wires, or within some boundaries allowing for 120 degree phase angles, then the ground configuration can be known. If the EGC does not seem to relate to the power it gets, it should enter "disabled" operation mode (e.g. shutdown to just enough to display a status of the problem).
| It would probably be easier to leave the grounding to a 'user configurable' | setting. Have the user configure the output inverter ground connection | before installing the unit and leave it connected all the time. Then when | you switch from bypass to UPS output, there is no ground-shifting either.
Assumes too much from the user. Of course, I'mt not expect 100% idiot proof. There will always be radical idiots. It should just be "plug and play" in any 200V to 240V system world wide.
--
|WARNING: Due to extreme spam, googlegroups.com is blocked. Due to ignorance |
| by the abuse department, bellsouth.net is blocked. If you post to |
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
** FUCK OFF
YOU ASININE RADIO HAM IMBECILE !!!
..... Phi l
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.