Another hot tub wiring question (long) - best 3-phase service?

This rather long question is directed to electricians and engineers who are familiar with wiring and service installation requirements for a commercial hot tub business. The question is what is the best type of service class (voltage - # of phases) for a U.S. installation with 6 or more hot tub units.
With the exception of some 120 V. installations, virtually all of the residential hot tub units in the USA specify single-phase 240 V. four wire service (usually at 50 or 60A). This consists of two hot wires, a neutral, and a safety ground wire. The code requirements specify a hefty Ground Fault Circuit Interupter for each unit.
The problem I am encountering in a commercial installation is as follows. When combining six or more of these units, it is presumably desirable to connect to a commercial size three phase service entrance in such a way that all of these large current-drawing pump motors and electric heaters present a balanced load to the incoming feeders. The problem is that even the commercial tubs require 240 volt single phase 4 wire service + the GFCI. Hot tub pump motors that operate on 208V single or 3-phase are non-standard inventory and difficult to obtain. The same applies to the factory standard electric heaters which expect 240 volts and take a performance hit if the only voltage available is 208.
Given the standard services available in the USA (Edison split-phase 120/240, 3 phase wye 120/208, 3 phase High Leg Delta 120/240, and others, etc.) each present difficulties to these requirements.
Standard Edison split-phase 120/240V service would require wiring all 6 hot tubs to one (or two if Delta primary) service phases and create a highly imbalanced situation as far as maximum current draw. I suspect that the power company would not be happy with this if the entire building were loaded onto one phase.
3 phase wye 120/208 service does not match with the required 240V. that the tubs specify for pumps and heaters.
3 phase High Leg Delta 120/240V service provides for the proper 240 voltage, but the problem is that two sides of the Delta do not have a center-tapped neutral, thus causing difficulties with the code requirements for a neutral and the GFCI.
My guess would be that the 3 phase wye 120/208 with three boost autotransformers to step up the 208 to 240 volts would be the best way to go, but I'm not an expert and certainly not as familiar with the code as some of the participants in this newgroup. The goal is a safe installation that meets and exceeds the NEC requirements at the lowest possible cost.
I will be doing further investigations and talking to my electrical contractor and the power company for advice, but I was wondering if anyone out there had all ready encountered a similar installation and perhaps could describe how they did it.
Incidently, this seems to be the one example I've found where the European System (High Leg at 240 V., current carrying neutral, and safety ground) seems to present an advantage over the US 3 wire split phase system. If the former were allowed in the USA, we would specify wye connected transformers with 240 volt secondaries, use Euro style tubs with their own GFCI's and evenly distribute the load on each phase. I'm not sure if the US code would permit this, however....
Beachcomber

snipped-for-privacy@nospam.xyz (Beachcomber) wrote in message
  Quoted Text Here  
<snipped>
My first inclination would be to ask why you want to install hot tubs rated for residential use in a commercial installation?
Presumably you have 120/208/4w coming in already. I'd be inclined to buy transformers and make 240v/3w (i.e. - three 208 to 240 center tap). They won't be real cheap or something you can get at home depot tomorrow, but acme or some other xfmr mfr can make them up pretty quick these days.
I am not a big fan of buck/boost xfmrs, having seen way too many of them applied improperly, just to save a few bucks.

  Quoted Text Here  
I thought I stated it earlier, but perhaps I did not make myself clear., these would be commercial units. Most, if not all of the commercial units available would still be rated for 240V. (same as residential).
Beachcomber

| |><snipped> |> |>My first inclination would be to ask why you want to install hot tubs |>rated for residential use in a commercial installation? |> |> | I thought I stated it earlier, but perhaps I did not make myself | clear., these would be commercial units. Most, if not all of the | commercial units available would still be rated for 240V. (same as | residential).
I'll add to this that my bad experiences with air conditioners were with commercial units rated for 240 volts. Apparently they are made with a motor option, either 240 volt 1-phase or 208 volt 3-phase. But you can't just switch between them by wiring; it's a whole motor changeout. The problem at the place I used to work where they had this issue was that even though the building had 3 phase power, the wiring out to the rooftop for the A/C units did not allow for 3 phase. They hooked the 240 volt single phase motors to 208 volts, and under heavy loads in the summer they were burning out. Eventually the A/C guys tracked down a 208 volt single phase motor that fit and the problems were solved.
Just because things are commercial units, don't assume three phase.
My points are:
1. Commercial stuff does come made for either 240/120 volt single phase or 208/120 volt three phase (or other voltages). Single phase stuff on voltages only seen with three phase service is rare, if it makes sense to make it in genuine three phase (lighting, for example, does come in 277 volts for places with 480/277 since it doesn't make sense to make 3 phase light bulbs).
2. Stuff made for 240 volts does not work well on 208 volts (such as ovens and heaters, due to the slower warmup times), and in some cases causes problems and breakdowns (motors drawing heavier current at lower voltage causing greater heat buildup and damage).
3. Stuff made for 240 volts often includes 120 volt components due to the usual expected center tapped wiring, so a neutral in the middle is still needed.
4. 240 volta delta three phase service can have a center tap on one winding to provide 240/120 volt service, but the load is limited to 5% of the overall capacity (usable in many industrial shops to power big motors at 240 volts three phase, and several office lights at 120 volts.
If we had never allowed a center tapped single phase service to be offered, then we would not have to deal with equipment designed for that in cases where we have to take three phase power for whatever reason. In much of Europe, three phase is available even for residential, and where you get 2 or even 1 leg of it, the voltages match the design of available products. With 230 volts there's less need for the next higher voltage, and when it is needed, it's always 1.732 times as much.
There, I get my "Europe did electricity right" plug in, yet again :-)

Where does this 5% limit come from, and why?
Thanks in advance.
wrote:> 4. 240 volta delta three phase service can have a center tap on one winding
  Quoted Text Here  

| Where does this 5% limit come from, and why?
That's a good question. Where I got it from is transformer manufacturer specifications. It seems to be very consistent: every one that has that much detail has exactly the same 5% figure. As to why, I don't know. I would think there is some limit, but my first guess would have been not less than 33%. It could also be as simple as that's the capacity they put on that center tap because that's what the market has wanted.
If I hang a 240 to 240/120CT transformer across 2 legs of 240 delta, what would be the limitation there under various 3 phase loads from 0% to 100%? Of course I expect 0% for what I could draw from that tap when the 3 phase load is 100%. But what I can do with less, and does that translate into the load I can put on the delta's center tap?
| Thanks in advance. | |
| wrote: | |> 4. 240 volta delta three phase service can have a center tap on one | winding |> to provide 240/120 volt service, but the load is limited to 5% of the |> overall capacity ....................... | |

  Quoted Text Here  
That may be the reason we see so much "Open Vee" center tapped delta. They will upsize the center tapped transformer to handle any extra 120v loads. I am not sure what the limit is but it is certainly a lot more than 5%. I had a data center running on an "Open Vee" and they had a lot of 120v office stuff along with the 3p computer system. They had one real fat transformer and a smaller one.

| I am not a big fan of buck/boost xfmrs, having seen way too many of | them applied improperly, just to save a few bucks.
Neither am I. And they are even trickier with three phase.
"Save a buck and get a boost in the rear".

  Quoted Text Here  
That could be said about almost any installation of any product!
The problem you've seen is crap installations not the Buck-Boost Xfmr proper. Properly specified, and installed, a Buck-Boost xfmr installation is neither complex or unreliable. Just the opposite is usually the case. Proper application can extend the life of the connected components due to operating within a proper voltage window.
There are ample sources to guide the proper application of a Buck-Boost application. A Google search is just one of many.
Louis-- ********************************************* Remove the two fish in address to respond
Louis

If you want to sell this in the US you had better use 240/120 volts 1-phase 3 wire or 208/120 volt 3-phase 4 wire. These are the standard voltages and phases. Try anything else and you are asking for trouble.

  Quoted Text Here  
Actually, the standard is 240/120 volts 1-phase 4 wire (Hot, Hot, Neutral, Safety Ground) for most residential "skid pak" hot tubs. (Some will run on 120V if that is all that is available). The code is rather stringent about the use of a GFCI for any residential installation. (Siemens makes 50A and 60A GFCI's for this purpose.
Agreed, three phase pumps and three phase hot tub heaters would be better, cheaper, more efficent, quieter, and probably last longer. The problem is that, as far as I can tell, the hot tub industry manufacturers do not stock them as standard equipment. Of course it is possible to build a custom system using 3-phase pumps and 3-phase DHW heaters are certainly available. There are increased costs and warranty issues for the custom construction option, however.
Beachcomber

  Quoted Text Here  
to
1-phase
are
In the trade it is called 240/120 volts 1-phase 3 wire and 208/120 volts 3-Phase 4 wire. The grounding conductor is not considered part of the system description. Just for your information.

  Quoted Text Here  
I am not sure that is really true if you are not simply using a skid pack. Pumps is pumps when it comes to warranty issues. A reputable company will back either one. A 3p unit is generally going to be a better unit than a consumer grade 1p pool pump but they are not that unusual in commercial setups and I thought that was what you were building. You will have less problems with a 3p pump motor since you don't have the "start" hardware.

Beachcomber wrote:
  Quoted Text Here  
If I understand what you are suggesting, this would give you a 240V 3phase Y. The problem here is that the line to neutral voltage will be about 138 volts and not suitable for any 120 volt loads within each unit.
You could use three 208 to 120-240V single phase transformers (not autotransformers) and feed three seperate single phase sub panels. But this is getting expensive.
You could also rewire the units to seperate the 120 volt loads from the 240 volt heater and motors. The 120 volt loads could be connected to the center tapped leg of a 240V delta and wouldn't produce significant imbalance. The problem here is that any UL cert. of these units would be voided, since it applies to the unit as built. This might work if the components are individually certified and the rewired units could be approved as 'field assembled' per another poster's suggestion.
  Quoted Text Here  

I think that this issue is one where the simplest, most direct answer is also the best - Just use a single phase 208 to 120/240 transformer on each unit. No modifications to the building wiring, no modification to the tub, no issues with "floating" neutrals.
  Quoted Text Here  

| I think that this issue is one where the simplest, most direct answer is | also the best - Just use a single phase 208 to 120/240 transformer on each | unit. No modifications to the building wiring, no modification to the tub, | no issues with "floating" neutrals.
That depends on if his building is already wired for what configuration of power. I've seen the effect of switching things from 240/120CT to 208Y/120. If the building is already wired up for 208Y/120, then it is probably cost effective to just leave it that way. He mentioned 6 hot tubs. Whether the best solution would now be 6 transformers (1 for each tub) or 3 (1 for two tubs, but still keeping the load balanced) would be an issue to consider. But if the building is yet to be wired (or rewired by remodeling) I would tend to go with 3x240/120CT just so I have true 240 available where needed. There are other factors that could sway the decision, such as how much load there is, if there is any real 3 phase load, and if I am getting power in at some other voltage (e.g. 416Y/240 or 480Y/277) that forces a transformer to be involved for everything, anyway. If there have to be whole-building transformers, that makes going with 3x240/120CT an easier decision for me. Then I have to weigh the cost of using a normal 3 phase transformer to get 208Y/120 plus the costs of the individual transformers for the 240/120 loads, vs. the cost of 3 single phase 480x240 to 240/120 transformers and no more needed for loads that do want real 240 or 240/120CT.

  Quoted Text Here  
Thanks to all for a good discussion. The spa building is still in the planning stages (without existing service) so that will figure into the specifications. Much useful input was given from these discussions.
Transformers are interesting, seemingly simply devices... In reality however, they are quite complex, like a game of chess, in terms of a theoretical analysis of the basis underlying principals.
For those interested in a little bit of history...
The inventors of the transformer (Lucien Gaulard & John Gibbs) experienced years of difficulty with their invention because they kept on hooking up their circuits with the primaries in series instead of parallel. The end result was that they could transmit power for long distances, but the practicality of the invention suffered due to poor load regulation. The parallel primary connection (for multiple transformers) that seems so obvious to us today, was not perfected until a theoretical understanding was applied by men like Tesla, Stanley, and Westinghouse.
Lucien Gaulard was born in Paris in 1850. Gaulard was a French scientist, primarily interested in chemistry of explosives, but later he shifted his interests to electrotechnics. He developed a thermochemic battery and he is particularly known for his work with induction coils (transformers). In 1882, Goulard and his English colleague Gibbs patented a system of distributing power using alternating current and two-coil induction devices. They used devices (then known as secondary generators) of the Ruhmkorff type in the first alternating current distribution system and had a 1:1 ratio and were used with their primaries in series.
A power transformer developed by Lucien Gaulard and John Gibbs was demonstrated in London in 1881, and attracted the interest of Westinghouse. Transformers were nothing new, but the Gaulard-Gibbs design was one of the first that could handle large amounts of power and promised to be easy to manufacture. In 1885, Westinghouse imported a number of Gaulard-Gibbs transformers and a Siemens AC generator to begin experimenting with AC networks in Pittsburgh. In hands of Westinghouse the transformers received really practical implementations.
Gaulard's transformers were successfully presented in 1884 on the international exhibition in Turin. The farthest lamp fed, on the Torino-Lanzo railway line, was at 40 km distance from the 2,000 V generator with 133 Hz frequency. The series connection led to unsatisfactory regulation unless all the transformers were equally loaded. However, these transformers were in use until 1912.
Gaulard was ingenious, but unlucky inventor. During his life his work was not recognized in France. Gaulard fell in depression state and seems he went mad, was sheltered in clinic where 1888 died the 26 November. Now Gaulard fame is memorized with tombstone in the railway station of Lanzo, there is also a street with his name in Paris.
This man was the co-inventor of one of the world's most revolutionary inventions that gave the world the ability to transmit electric power for long distances, but for some reason he just couldn't get the connections correct.
Beachcomber
Reference: (excerpts from) http://chem.ch.huji.ac.il/~eugeniik/history/gaulard.html

| Thanks to all for a good discussion. The spa building is still in the | planning stages (without existing service) so that will figure into | the specifications. Much useful input was given from these | discussions.
That will mean you have a choice in the way you do this. You could go with getting 208Y/120 service and smaller transformers to take 208 volts up to 240/120CT. Or you could go with that or other three phase voltage and use transformers for your full capacity. With that choice you can either go with 240/120CT for the whole building, or do 208Y/120 and more transformers to get 240/120CT just for the tubs. You will need to see what else you might have that needs, or works better, on 240/120, and evaluate the cost of all your options.
| The inventors of the transformer (Lucien Gaulard & John Gibbs) | experienced years of difficulty with their invention because they kept | on hooking up their circuits with the primaries in series instead of | parallel. The end result was that they could transmit power for long | distances, but the practicality of the invention suffered due to poor | load regulation. The parallel primary connection (for multiple | transformers) that seems so obvious to us today, was not perfected | until a theoretical understanding was applied by men like Tesla, | Stanley, and Westinghouse.
You mean multiple transformers feeding different loads where the primaries for each are in series? Yeah, that would be bad.
You can get transformers would double primaries which can be wired either in series or parallel. One rated 480x240 would handle 480 volts when wired in series and 240 volts when wired in parallel, with the same kVA capacity.
If you get 480Y/277 power, you could get one 3 phase transformer to take 480 volts in to a delta winding, with a secondary of 208Y/120 (three 120 volt windings in wye configuration). Then instead of doing 208 volts to 240/120CT, you could tap off the 480 volt side with the tub transformers and use the cheaper (because they are more common) 480x240 to 240/120 transformers that are out there.
| Lucien Gaulard was born in Paris in 1850. Gaulard was a French | scientist, primarily interested in chemistry of explosives, but later | he shifted his interests to electrotechnics. He developed a | thermochemic battery and he is particularly known for his work with | induction coils (transformers). In 1882, Goulard and his English | colleague Gibbs patented a system of distributing power using | alternating current and two-coil induction devices. They used devices | (then known as secondary generators) of the Ruhmkorff type in the | first alternating current distribution system and had a 1:1 ratio and | were used with their primaries in series.
Ah, I see. They probably didn't have a real grasp on the designing for particular voltages, capacities, saturation, heat rise, etc, as we (not me personally, but those who do make them) do today.
| Gaulard's transformers were successfully presented in 1884 on the | international exhibition in Turin. The farthest lamp fed, on the | Torino-Lanzo railway line, was at 40 km distance from the 2,000 V | generator with 133 Hz frequency. The series connection led to | unsatisfactory regulation unless all the transformers were equally | loaded. However, these transformers were in use until 1912.
Interesting frequency. I wonder how they picked that one.
| Gaulard was ingenious, but unlucky inventor. During his life his work | was not recognized in France. Gaulard fell in depression state and | seems he went mad, was sheltered in clinic where 1888 died the 26 | November. Now Gaulard fame is memorized with tombstone in the railway | station of Lanzo, there is also a street with his name in Paris. | | This man was the co-inventor of one of the world's most revolutionary | inventions that gave the world the ability to transmit electric power | for long distances, but for some reason he just couldn't get the | connections correct.
Perhaps he was victim to the same kind of misunderstanding I once had where transformers didn't make sense (despite knowing they should). It seemed to me back then that having _more_ windings in the primary would create _more_ flux, and that should produce more "secondary generated" power. Of course the reality is more winding is more impedance and less current. You don't want to tie them all in series with each other on different loads, or else you are making current transformers out of them (and pathetic ones at that).

  Quoted Text Here  
Thanks for the suggestions. One question though... If I did as you suggest above with three single phase 120-240 transformers, I have concerns about the derived neutrals. Since the primaries would be coming from different phases, it would seem to me that there would be a potential difference between the neutrals of each sub panel if each neutral were connected to the center-tap of the secondaries. In other words, is a neutral truly a neutral if it is not grounded somewhere? If everything were isolated, it would be as though the appliance were operating through a full isolation transformer and would lose the benefits of a safety ground? Does the code allow this?
Hence I thought that by using autotransformers, the original continuity of the (grounded) neutral at the center of the wye would be preserved.
I hope I am expressing this clearly. It seems to be a complex issue.
Beachcomber

|>You could use three 208 to 120-240V single phase transformers (not |>autotransformers) and feed three seperate single phase sub panels. But |>this is getting expensive. |> | | | Thanks for the suggestions. One question though... If I did as you | suggest above with three single phase 120-240 transformers, I have | concerns about the derived neutrals. Since the primaries would be | coming from different phases, it would seem to me that there would be | a potential difference between the neutrals of each sub panel if each | neutral were connected to the center-tap of the secondaries. In other | words, is a neutral truly a neutral if it is not grounded somewhere? | If everything were isolated, it would be as though the appliance were | operating through a full isolation transformer and would lose the | benefits of a safety ground? Does the code allow this?
There is no issue regarding derived neutrals. You don't derive them; you just connect them together. This is what I believe is suggested:
[be sure to display with a fixed width font, such as Courier]
A-----*------------------------------------------------------* B-----|--------------*----* | C-----|--------------|----|--------------*----* | N--* | | | | | | | | 208 | | 208 | | 208 | | \/\/\/\/\/\/\/ \/\/\/\/\/\/\/ \/\/\/\/\/\/\/ | ================= ================= ================ | /\/\/\/\ /\/\/\/\ /\/\/\/\ /\/\/\/\ /\/\/\/\ /\/\/\/\ | | 120 | 120 | | 120 | 120 | | 120 | 120 | | *--------|--------|-|--------|--------|-|--------|--------|-- X1 | | | *--------|--------|-|--------|--------|-- Y1 | | | | | *--------|--------|-- Z1 *----------*--------|----------*--------|----------*--------|-- N *-------------------|-------------------|-- X2 *-------------------|-- Y2 *-- Z2
| Hence I thought that by using autotransformers, the original | continuity of the (grounded) neutral at the center of the wye would be | preserved.
Normally an autotransformer used for boost (or buck) is done with small increments of 12 to 32 volts of change. But technically it could do a larger change. Perhaps what you meant is this:
A----*------------------------------------------------------------ X1 B----|-------------------*---------------------------------------- Y1 C----|-------------------|-------------------*-------------------- Z1 N----|--------*----------|--------*----------|--------*----------- N | | *-|--------|----------|--------|----------- X2 | | | | | *-|--------|----------- Y2 | | | | | | | | *-- Z2 | 120 | 120 | | 120 | 120 | | 120 | 120 | \/\/\/\/ \/\/\/\/ \/\/\/\/ \/\/\/\/ \/\/\/\/ \/\/\/\/ ================= ================= ================ Instead of trying to extend each of the three legs longer (to 139 volts to get 240 between legs), what you'd do in this case is extend a "new leg" in the opposite direction and derive a 120-0-120 configuration from a 0-120 configuration, times three for all phases (splitting hot tub loads across the three phases equally). Someone who knows more about this than I do will need to make sure the ratings are right, and check to make sure this will meet code, especially in the wet situation. Finding such a transformer would be another matter. Something with identical 120 volt windings on primary and secondary might be usable.
I'd still prefer the isolation from the first case, though the second looks like it could work. Either way you get a 6-star output:
Y1 * * Z1 \ / \ / X1 *----N----* X2 / \ / \ X2 * * Y2