Switched-mode power-supply

[<Magnus Pihl>]

Is it possible to make a totally switched-mode power supply, so I don't need to use my 20kg original PSU? I mean to make the filament transformer part a switched-mode too. I want to drastically loose weight on the PSU!!!<p>Any schematics available on this or other models? I have seen that "Reliant 1000M" uses a switched-mode type.<p>Best regards,
Magnus Pihl

[wler]

Hi,<p>I assume you mean a PS for an argon laser ?
Depending on how much current it takes, it is relatively easy possible or difficult; I would
say until 30A it is feasible, perhaps more. I am
just finishing a SMPS project which works fine, and spent some time to set up a hopefully useful description, see <a href=http://perso.wanadoo.fr/redlum.xohp/swi ... ml>here</a>.

[<Magnus Pihl>]

Hi,<p>Interesting to see that others are curious to make a SMPSU to their argon laser too. I have a 230mW multiline polarised argon laser.<p>Are your SMPSU type a "buck" regulator? If so, that means it is not isolated from the mains. I have done calculation and experiments with ferrite core transformers. They are isolated from the mains. Therefore, I use no inductor, but a lightweight transformer.<p>I know how to design a isolated transformer, but what bothers is these questions:<p>1. The argon laser dissipates about 12A on the 230VAC line at maximum load, they say. This is 2760W. Is this the "tube current", which flows in the plasma between the filament (cathode) and the anode, or only in the filament?<p>2. The "tube voltage" is 100VDC, people say. If this is the fact, and 2760W dissipates over anode-cathode, the "tube current" must be 2760W/100V = 276A. Now, this can't be true!! The diodebridge can only handle 25A.<p>3. When are the "boost" voltage needed? During startup-ionisation or all the time?<p>Best regards,
Magnus

[wler]

Hi !<p>>Are your SMPSU type a "buck" regulator?
yepp !<p>>If so, that means it is not isolated from the mains.
True.. but if you touch an isolated one there is not such a big difference, due to the capacitors.<p>>I have done calculation and experiments with ferrite core transformers. They are isolated from the mains. Therefore, I use no inductor, but a lightweight transformer.<p>Well I admire this, this is much more complicated to handle than a buck regulator; all the
subtleties with "leakage inductance" and
voltage spikes on the drains. I gave up on ferrite and turned to certain powder cores, as described. How big is your transformer and of what is the ferrite material exactly ?<p>>I know how to design a isolated transformer, but what bothers is these questions:<p>>1. The argon laser dissipates about 12A on the 230VAC line at maximum load, they say. This is 2760W. Is this the "tube current", which flows in the plasma between the filament (cathode) and the
anode, or only in the filament?<p>Well this depends on the tube voltage. You say below it is 100V, so at 12A this makes 1200W
which goes in the tube via the anode. Add perhaps
60W to the filament, too. I don't think that under
these circumstances the PS would draw 12A too,
rather something like 1/2 of it, depending on
the efficiency of the PS. A SMPS would dissipate
only a few tens of Watts in addition. <p>Only if you run a linear mode regulator it would draw 12A also at 230V and this would create an enormous heat, rougly 2760W -1200 =1560W in your PS, which is very hard to handle. Better is to use
a stepdown transformer, which is what most ppl use.<p>2. The "tube voltage" is 100VDC, people say. If this is the fact, and 2760W dissipates over anode-cathode, the "tube current" must be 2760W/100V = 276A. Now, this can't be true!! The diodebridge can only handle 25A.<p>Well, 2760/100 =27.6A. But it's more like 1/2 of this, see above !<p>3. When are the "boost" voltage needed? During startup-ionisation or all the time?<p>Only for start-up.<p>Again, I am eager to hear about your transformer
experiences, this was the biggest headache for me.<p>Best,
Wolfgang

[magnus pihl]

Hi,<p>Sorry about the mistake in the calculation. I saw the mistake just when I pressed "post reply", but I was unable to delete/edit the reply.<p>I don't have a working SMPSU for my argon laser yet. Sorry if I gave you that impression. I use MediaLas PSA12 now. That is a 20kg piece, and I don't like the weight of it. I assume the 230VAC to 70VAC (100/sqrt2) transformation is the part where most of the weight are.<p>I do have designed SMPSU and know how to calculate the core. I will use a ETD core in the efficient push-pull mode. I am skilled to deal with the EMC problems. <p>To adjust tube current, I assume I have to reduce the anode voltage, right? If so, that can simply be made by slightly adjusting the PWM by a servo loop with tube current monitoring, -not by frying off power by transistors, like PSA12 do.<p>I will simply remove the "15A Variac" and recifitying part by a AC/DC converter. <p>What do you think?<p>Best regards,
Magnus Pihl

[ralf-k]

Hi!<p>I have build a SMPS too. I use a ETD 59 core. But it run not so good. So I use a ring core (this tip is from Wler). Now it run better.
Photo of the home build (not runing) SMPS.<p>I know, my English is very bad [img]images/icons/frown.gif"%20border="0[/img] [img]images/icons/mad.gif"%20border="0[/img] [img]images/icons/frown.gif"%20border="0[/img] <img src="graemlins/cry.gif" border="0" alt="[Cry]" />

[magnus pihl]

Hello Ralf,<p>What are the problem with your SMPSU?
I have some questions about it:<p>1. Are the "filament transformer" outside the PCB? It isn't in the ETD49 core?<p>2. Do you use the ETD59 in "push-pull" (primary has three terminals and at least two transistors) mode or "feed-forward" (primary has two terminals and one MOSFET that pull down to ground) mode. The last one only gives 50% effiecency of the core.<p>3. What switching frequency do you use? You need about 250kHz, I think.<p>3. I see no "demagnitizing diodes" across the primary winding(s). This is necessary for a sucessful operation.<p>4. How do you adjust "tube current"? By PWM to the primary winding?<p>Best regards,
Magnus Pihl

[wler]

Hi Magnus !<p>(BTW where are you from ?)<p>>To adjust tube current, I assume I have to reduce >the anode voltage, right?
Yupp
>If so, that can simply be made by slightly >adjusting the PWM by a servo loop with tube >current monitoring, -not by frying off power by >transistors, like PSA12 do.
Exactly - I use a SG2525 for that, see the circuit; it regulates very well.<p>>I will simply remove the "15A Variac" and >recifitying part by a AC/DC converter.
Not sure - do you mean that there
is a variable transfomer, a variac, in your PS ?
That must be heavy indeed if it can take 12A.
So do you mean to leave the PS as-is but change only the input transformer and rectifier ? My comment would be that if you change that part of the circuit, you don't need the remaining linear regulator any more, because its role can be
played by the SMPS as well.<p>Best,
Wolfgang

[magnus pihl]

Hi,<p>Well, no I don't use a variac. I had the schematics of a linear PSU of the ALC60 from www.laserfaq.com in front of me, when I wrote the reply. But the PSA12 use a 15kg transformer. That is just like a variac, but with no wiper. So, -it still got the weight! <p>I mean that I will use the technique as other power-supplies use, like boost and so on, but I will replace the AC/DC part. In the end, I will have two transformers. One ferrite core and one 50Hz iron-core transformer for the filament. This
should be quite light-weight! The linear regulator which fries off the power, will be eliminated of course, as the PWM do the rest.<p>I tried to find the datasheet of the SG2525 on the internet, without sucess. Can you send me a link to the datasheet? Does it drive "feed-forward" (3-terminal primary) or "push-pull" (2-terminal primary)?<p>I live in Stockholm, Sweden.<p>Best regards,
Magnus

[ralf-k]

Hi!<p>Look here.
Thats the datasheet from SG2525

[wler]

Hi Magnus,<p>Ralf has already linked it;
I just confirm that this IC indeed supports both
single ended and push-pull configurations.
I found however it is more useful to use a separate mosfet driver IC, as this gives more layout flexibility and allows for shorter
switching times. At any rate I have put links for
all relevant data sheets for my SMPS on my site.
Otherwise the best site for datasheets I have found so far, is http://www.freetradezone.com.

[magnus pihl]

BTW, how critical is the filament voltage? I have read that the voltage should be 3.2V. On another site I read 5V. <p>Magnus

[wler]

Well the filament voltage depends on the specific
tube you have and is typically between 2.5 and 3.3V.<p>As for the anode voltage, this is practically
independent of the current. The tube has a very
nonlinear characteristic, say takes 100V almost
no matter whether I=4A or 10A (that is, the "differential resistance" which measures how much the current changes if you change the voltage by very little, is very low). In other words, if you would change the voltage from say 100.0 V to 100.1 V, it would make a big difference in the current.<p>This is why the PS must be current regulated and not voltage regulated.

[magnus pihl]

To adjust the tube current, you adjust the PWM to the ferrite core. Question: Between which voltages (approx.) do you control the tube current?<p>If the PWM goes too low, I assume the ionazion disappear.<p>Best regards,
Magnus

[ralf-k]

Hi!<p>Look here. Thats the tube voltage and the current from a ALC 60:<p>2,5A...94,2V

3A.....94,2V

4A.....95,0V

5A.....96,7V

6A.....99,1V

7A....101,5V

8A....104,5V

9A....107,7V

10A...112,0V

11A...115,0V