Enterpise power supply + I70, possible?

[DMills]

Hi all,

Having just picked up a known good I70 head (Argon, multiline) for the right kind of money (Witnessed doing 5W, may get more when cleaned and tweaked if I am lucky), the question arises as to what to use as a power supply....

Now I got an old Coherent Enterprise thrown in with the deal that has a dead head, and I am thinking that that (fucking heavy) switchmode looks like it could hack the 35A that would be needed (And I **think** the Cathode voltage and current are the same). The nice thing about this approach is that it will run on single phase 240V.

The Enterprise looks (on a fairly quick look) like a thyristor input switcher (That huge 0.5mH choke), followed by a Mosfet passbank (Split across 3 cards, but this must run with a fairly low voltage drop (or it could be some sort of switchmode) as the heatsinking is not all that serious.


Anyone (Steve?) know what the odds of building a FrankenGas are with this combination?

A schematic (or even a manual) for the Enterprise would be really helpful (that head has altogether too much electronics in it) if anyone has such, I doubt very much that Coherent will be forthcoming with relevant info.

Someone has obviously been in the power supply as the three (current regulator?) boards were rattling around loose, so the first thing is to put them back, figure out the interlocks and see if I can get the thing to load up into a couple of buckets of salt water.....

I am thinking that moving the 'tronics from the Enterprise head and transplanting the loom onto the I70 would be a good first step once I have power available from the supply.

Edit The current regulator turns out to itself be switchmode!
Gaa, this thing is complicated (And it really does have FAR too much logic for its own good.
I am getting tempted to just design a simple PWM chopper kind of current regulator and have the input thyristor bridge maintain its DC at say Vtube + 15V or so, should allow for suitably small magnetics.

[krazer]

I am not familiar with the enterprise supply, but what I have seen from them they appear to be the 'smart' supplies that would normally be associated with a nice DPSS head, the kind that will do little more than turn on without a matching set of EEPROMs and whatnot.

Hopefully you can hack it together, argons are one of the simpler things to drive from a switchmode PSU all things considered.

[buffo]

Having just picked up a known good I70 head (Argon, multiline) for the right kind of money (Witnessed doing 5W, may get more when cleaned and tweaked if I am lucky), the question arises as to what to use as a power supply....

Holy crap! 5 watts from an I-70? Either it's brand-spanking new, or you're one lucky bastard. That's a hot tube, buddy. Hang on to it.

As far as building a frankengas, I'll defer you to Steve Roberts, as I don't know anything about the Enterprise PSU's. However, I do have most of the schematics for the original Coherent medical I-60/I-70 3 phase power supply if you want. (And yeah, they're every bit as complex as you'd imagine...)

Adam

[mixedgas]

[QUOTE=krazer;144854]I am not familiar with the enterprise supply, but what I have seen from them they appear to be the 'smart' supplies that would normally be associated with a nice DPSS head, the kind that will do little more than turn on without a matching set of EEPROMs and whatnot.

Abandon all hope ye who enter here.... Smart supply, matches the tube I/V curve to a given head via eeprom. I've never seen a schematic for them anywhere, anyhow... I do have data for the 3XX series they are derived from. Even if you did get it to go, odds are you'd need a service prom to recal the supply.

I do have a I70 PSU for cheap if you want to pay shipping......

I think what you need is the I70 tube specs, that I can do, give me a few days, the weekend just started. Or go see SMOG, that 88 supply of his would just light up and run up the I70 tube.. Might newt get that much power, but it would light it.

Enterprise is 162.952% more complicated then the one you have Adam.... :-)

Hey Dmills, do shoot us a pic of the psu insides, maybe there will be some commonality,,

first thing I would do is bridge the 70 series tube into the enterprise supply as/is, you cant hurt it, and they usually take a few minutes to say they dont recognize the tube, if they are like the 3XX series. Actually you get about 20 minutes on the 3XX while it determines it does not like the fill pressure.

Steve

[DMills]

.
Abandon all hope ye who enter here.... Smart supply, matches the tube I/V curve to a given head via eeprom. I've never seen a schematic for them anywhere, anyhow... I do have data for the 3XX series they are derived from. Even if you did get it to go, odds are you'd need a service prom to recal the supply.

About what I figured when I saw that logic board in there, come back 265s all is forgiven!

Ahh well, I have the case, the FETs and the inductors, also the big raw DC section, time to try my hand at ion laser power supply design....

I do have a I70 PSU for cheap if you want to pay shipping......

Not from the 'states, I could probably source one locally by the time I had paid the customs and shipping.

I think what you need is the I70 tube specs, that I can do, give me a few days, the weekend just started. Or go see SMOG, that 88 supply of his would just light up and run up the I70 tube.. Might newt get that much power, but it would light it.

I do have a OEM supply for a 909 here that will manage 30 amps or so, which I could use to at least keep the pressure under control while I build something much dumber from the guts of the old Enterprise supply.

Hey Dmills, do shoot us a pic of the psu insides, maybe there will be some commonality.

Ok, will do, may be a day or so, as you say the weekend is yet young.

Thanks.

Regards, Dan.

[DMills]

Well, having drawn a partial schematic of one of the current control switchers, it looks like a custom (and very much simpler!) control board is in now way impossible, and might (I am still figuring out some of the control loops) not be that hard to come up with.

The basic topology is a thyristor input bridge feeding a choke and big cap bank (This could be improved upon by using an IGBT and power factor correction chip, but for now it will do) which looks like it is probably setup to generate a raw DC rail at maybe 2-300V or so, kind of hard to tell.

There are then 3 parallel current control cards that consist of a basic Buck topology current mode switcher (UC3843 controller, 24N50 fet, current transformer feedback), There is a lot of additional analogue stuff in this area that I am still trying to figure out.
The voltage loop feedback appears to come from off card so I am thinking that this is used to command tube voltage from the DC current sense board at the bottom of the stack.

There is an additional card at the bottom of the current control stack, purpose I am still trying to figure out, looks like DC current monitoring and some protection stuff but it is hard to tell without drawing a diagram.

So far this thing is nowhere near as scary as I suspected it might be (The original control logic has to die, but the power side is basically simple).

Just spotted something 'interesting' that suddenly makes this project MUCH less of a pain.....

The control card is basically split down the middle, with all the interface stuff (including eproms, PALs and all sorts of yuck) at the back and just 2 signals going forward and two back to the power section....
One going in each direction is optocoupled via a 4N37 (probably enable and power good, or similar so that should be easy to figure out), and in addition there is a current sensor output (as a return from the hall sensor on the lower power card) and a IL300 linear optocoupler that clearly MUST be the current control feedfoward (There isn't anything else) Finally there is a relay drive signal for the main contactor.

This means that the current control error amp must be buried in the icky logic, but that that plus the interlocks and startup sequence is probably all that is required to lobotomise this thing.

Just got to find somewhere to stick a bit of stripboard with the new control loop and timing on it....

I think the biggest pain now is going to be replacing the keyswitch (which is well buried and totally inaccessible from the back :-(

Game on!

Regards, Dan.