Power Supply

Anything goes here, but abusive posts will be deleted.

Re: Power Supply

Postby steve_qix » Thu Sep 02, 2010 2:18 pm

A lot of this is "good engineering practice" - and some of it is "Style" (style, being subjective!! - just like in programming).

It is probably more "style" rather than "G.E.P." to shut down as many circuits as possible when they are not in use. Power supplies, in particular, generate heat when in use - even when idling, and if they are powering up other circuits (like the modulator) more heat will be generated. The components will last longer if they are not left on all the time. It is safer - far reduced risk of fire, shocks, explosions (yes, I've had them all!). If you leave the room for 3 hours and forget to turn the rig off, if everything pretty much shuts down between transmissions - it's no big deal.

I tend to be very conservative in every design and implementation I do. Been in broadcasting and engineering too long :D After a while, you've seen it all !!

I'm usually careful, but one time I went away for a WEEK - came back and the rig had been on the WHOLE TIME. Fortunately, when in receive, pretty much everything that uses more than a few watts goes dormant. Things happen, even when you don't think they ever will !!!! One time I actually left the rig in transmit for *5 hours* at full power! What a test, even if not on purpose. :shock:

Cats, wives and children (and children's FRIENDS) tend to go where they're not supposed to go...

One thing, though - the pulse width modulator should absolutely NOT be left running under any circumstances whatsoever when the transmitter is not actually transmitting. The pulse train should be killed so the modulator is not switching. Operating the modulator/filter at any time, and not terminated in something resembling its design impedance value can and is hazzardous to your modulator and filter components. Furthermore, if a load is suddenly applied to an operating, unterminated pulse modulator, there can be a staggering amount of voltage present, stored in the output capacitor of the filter and dumped into a load that is probably not made to take it.

Again, as part of the modulator design, I made it so the charge pump will not come into operation unless the modulator is terminated. However, this is not supposed to function as the primary shutdown mechanism (or any shutdown mechanism at all, actually). It is there to prevent the modulator from operating at all if there is no load.

There are a lot of "belts and suspenders" in the design :D 8)

Anyway, it is somewhat style, somewhat safety and some just plain engineering practice.

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby K9ACT » Fri Sep 03, 2010 12:17 am

steve_qix wrote:A lot of this is "good engineering practice" - and some of it is "Style" (style, being subjective!! - just like in programming).


Thank you for the thoughtful response. Exactly what I would expect from a real engineer. Your stories of forgetting about rigs running says a lot. Something I never thought of from a design point of view, in spite of it having happened to me a few times but never with a big tube rig. They are too hard to forget about and unlike someone we all know and love, I haven't fallen asleep on the air yet.

Speaking as one who spent most of my career selling electronics to engineers, you can well imagine that my "style" is a bit different, to put it mildly. The important thing is that we both end up reaching our own goals.

There is just on point that I want to make sure I understand....

>One thing, though - the pulse width modulator should absolutely NOT be left running under any circumstances whatsoever when the transmitter is not actually transmitting. The pulse train should be killed so the modulator is not switching. Operating the modulator/filter at any time, and not terminated in something resembling its design impedance value can and is hazzardous to your modulator and filter components. Furthermore, if a load is suddenly applied to an operating, unterminated pulse modulator, there can be a staggering amount of voltage present, stored in the output capacitor of the filter and dumped into a load that is probably not made to take it.

At first glance this reads like a death sentence to my transmitter. However, on a closer read, it seems like my setup may be excepted from the "absolutely NOT" dictum.

I believe by "modulator" you mean the big fets (one in my case). As part of my keying circuit is to disconnect the output of the modulator from the load, there is no path for the HV other than the light bulb. The PWM generator is still running and the driver stage is running but they just seem to be pulses to nowhere causing no problems that I have seen yet. I ran the net for two hours today and left it on but unkeyed most of the rest of the day and it still works as of an hour ago.

I rather expected I might have receiver trash but turns out, I hear nothing.

So, making allowance for "style", am I missing something serious?

Jack











Again, as part of the modulator design, I made it so the charge pump will not come into operation unless the modulator is terminated. However, this is not supposed to function as the primary shutdown mechanism (or any shutdown mechanism at all, actually). It is there to prevent the modulator from operating at all if there is no load.

There are a lot of "belts and suspenders" in the design :D 8)

Anyway, it is somewhat style, somewhat safety and some just plain engineering practice.

Regards,

Steve
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby steve_qix » Sat Sep 04, 2010 11:14 pm

One question: does the transmitter (modulator) include the overload board?
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby K9ACT » Sun Sep 05, 2010 12:13 am

steve_qix wrote:One question: does the transmitter (modulator) include the overload board?


No but I have the board and bag of parts and just need some motivation to build it.

I abused it again today for hours and the only "problem" I noticed was that when I over modulate it, my bleeder light bulb dims a lot more than usual, the power drops to about 50% and (I think) the FET eventually fails. I had to replace the FET but have had no problems since I started using this as a danger warning. It happens as the modulation approaches 200% so if I keep it down to about 150% it seems pretty happy.

I ran my CQ robot for 60 minutes two different times today. That's 30 seconds calling and 5 seconds unkeyed with modulation at 150%. Nothing got above 80F in the modulator and the RF Fets got to about 110F. The mod has a small fan because the heat sinks are minimal but there is no fan on the RF deck. I will probably add one if I go for the 200W version.

BTW, George WA9UJR is moving right along and so we should have two of these in the area soon.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby steve_qix » Sun Sep 05, 2010 11:43 am

What FET fails? You may have a problem. No FETs should ever fail.

I guess the upshot of all of this is - if you want to have ongoing or occasional failures (and sometimes catastrophic failures) what you have built is fine.

The design AS DESIGNED and documented is such that you should never have a failure of any kind, other than those caused by otherwise defective components, electrolytics wearing out (after many years), etc.

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby K9ACT » Sun Sep 05, 2010 11:38 pm

steve_qix wrote:What FET fails? You may have a problem. No FETs should ever fail.


It was the FET in the modulator that failed and I am not too discommoded over losing a $2 device as part of my learning experience.

To get back to your style idea, my joy in ham radio is having my signal represent me and the more I contribute to that signal, the more joy I get. Clearly for me, there is no joy in talking on a ricebox because I have contributed nothing but my money. I had been off the air for some 20 years because there seemed little alternative to SSB and rice boxes.

Re-discovering AM has provided the motivation I needed to get back on the air and I have spent the past few years building AM equipment and using it. I am not smart enough to design the sort of stuff you are involved in nor am I likely to improve on it.

However, I am no stranger to the KISS notion and simplifying a design is as close as I can get to designing it. If I can get by without a whole bunch of stuff, I am much more likely to enjoy building and using it. If it costs me a FET now and then, it gives me something to do... that's what homebrewing is all about. If I wanted plug and play, I would have bought a DYY rig.

Tonight I got bold and fired it up on 3705 and had a QSO with KB3WL. The novelty to me was having to do nothing but change the VFO and touch up the antenna tuner. Broad band stuff still boggles my brain.

Great fun,

Jack
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby steve_qix » Wed Sep 08, 2010 7:42 am

Ok....

No FET should ever fail, particularly in the modulator. There are some 50 of those boards operating in the field, and we don't see failures.

There is some implementation problem that is causing this failure. A single modulator MOSFET will easily modulate a 250 watt (carrier) transmitter to 200% positive, and those boards are used regularly to modulate KW+ transmitters (with 5 MOSFETs and dampers installed).

A number of factors could be causing problems.

Possibility 1) The PWM filter uses non-air (iron, ferrite, etc.) core inductors, and they are saturating. Unless specific core materials that are designed for PWM filter inductors are used, the cores WILL saturate and this will definitely cause failures. If the inductors or cores get warm, that's a pretty good indication of what's happening. This saturation problem will not occur with air core inductors. Saturating inductors in PWM filters will not only cause failures, but will negate the filter, and PWM components will appear in the output. The best non-air (air is the best) core materials for PWM filters are High Flux, Mega Flux and X-Flux. Most other materials will *NOT* withstand the DC component present in all pulse width modulator filters, and will quickly saturate as the DC increases. There are no cheap core materials at all that will work in PWM filters, and specialty materials (such as X-Flux) are specifically designed for PWM filters where high DC bias is present. There is no way around this. The proper core material (or air) must be used.

Possibility 2) There is something in the load that is non-linear, pulling unsually high current from the modulator under high modulation conditions. The overload board would protect the modulator in this case.

Possibility 3) The output waveform is not returning to 0V (or close to it), and the charge pump is not being brought into operation. This would cause the DC to fall off, and the modulator could be operating in a linear fashion.

Anyway, those are some possibilities. Check it out and let us know.

Device failures are not a normal part of solid state AM life.

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby K9ACT » Wed Sep 08, 2010 8:53 am

steve_qix wrote:Ok....

Device failures are not a normal part of solid state AM life.



That's interesting because most of the on air talk a few years ago was about replacing FETS. Sounds like things have changed since you came up with your boards.

I must apologize for failing to recognize and mention a "detail" which probably caused the failure. I mentioned it yesterday on the NTF and was immediately told "DON'T DO THAT!"

In the process of learning and "trouble" shooting, I found "after" a serious malfunction that the scope ground lead was hot enough to soften the insulation and the whole scope cable was warm. I seem to have confused cause and effect when I found that the FET was shorted.

This is one of those "details" obvious to an engineer but went right by me. Apparently you do not want to connect the scope ground to the floating ground.

> A single modulator MOSFET will easily modulate a 250 watt (carrier) transmitter to 200% positive,

So I should be all set to go when I upgrade to 200W.

> This saturation problem will not occur with air core inductors.

My first inductor is air on the modulator. Do the same rules apply to the RF deck?

I sort of put the upgrade on hold pending resolution of the problem but now that I know it was me, I will get back to it.

One thing that is not obvious though is the fact that going to 200W without increasing the HV might offer little in peak power output improvement. 200% mod at 100W is the same as 100% at 200W if I understand this correctly.

Gotta do it though because this is working so well now it's getting boring and time to blow something up again.

Js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby steve_qix » Wed Sep 08, 2010 9:23 am

K9ACT wrote:
steve_qix wrote:Ok....

Device failures are not a normal part of solid state AM life.



That's interesting because most of the on air talk a few years ago was about replacing FETS. Sounds like things have changed since you came up with your boards.



Only people who had implementation problems or bad components (usually bad shunt capacitors - doorknobs are BAD in this application) had problems. I haven't replaced a FET in any of my transmitters since 2002 (and I have a lot of transmitters!). I have fixed a few misconfigured or badly implemented pieces of equipment over the years.

A combination of digital gate drive, multilayer ceramic capacitors for shunts, and the modulator boards have pretty much flushed out any problems. Occasionally I run into an anomalous implementation somewhere - scratch my head for a bit, and generally find the solution. It's interesting to see what people come up with and the problems that can be created (including yours!). All part of QC. If I see a problem occur more than a couple of times, I realize that it needs to be well documented (and one hopes that people READ it :P

Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby kf1z » Wed Sep 08, 2010 9:00 pm

K9ACT wrote:One thing that is not obvious though is the fact that going to 200W without increasing the HV might offer little in peak power output improvement. 200% mod at 100W is the same as 100% at 200W if I understand this correctly.


Js


You'd be increasing the current when you "go to 200 watts". But the voltage would be the same.
(though if you haven't already, it is highly reccomended that your HV should be 120-130vdc)

So if, for example, your 100 watts is derived from 40 volts at carrier @ 2.5 amps
and
NOW you double the number of mosfets....in the RF amplifier.
your current will increase due to the LOWERED resistance (On-state resistance of the FETS, and output tuning)
You still have 40 volts dc at carrier but now 5 amps thus, 200 watts.

So you still have the same carrier to peak voltage ratio you had before, so the same modulation levels can be realized without changing the voltage.
(obviously, the power supply must be capable of handling the increase in current)

In a class-e amplifier, the tuning of the output circuit would change...
I BELIEVE you change configuration... thus tuning.. of the Class-D output stage as well.
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Re: Power Supply

Postby K9ACT » Wed Sep 08, 2010 11:41 pm

steve_qix wrote:
Only people who had implementation problems or bad components (usually bad shunt capacitors - doorknobs are BAD in this application) had problems.


Glad you mentioned that as I am getting by with ceramics but might want to use something else for a bigger version and doorknobs come to mind.

This also brings up another problem I have. I don't know what a shunt capacitor is but I know I have seen it on your web site but I can not find it.

If you are open to suggestions, you might want to take a look at my home page for an idea on a simple search engine.
You simply plug in a key word like "shunt capacitor" and it will find a link to the page or pages. It's just a Google thing that I stealthafied so it looks like it's my own until you get the search results but you just ignore the Google stuff. I am not a programmer but I am good at cutting, pasting and "creative editing".

If you are interested, I can send you the code and you can do your own "creative editing".

My home page is... http://schmidling.com/

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby K9ACT » Thu Sep 09, 2010 12:08 am

kf1z wrote:
You'd be increasing the current when you "go to 200 watts". But the voltage would be the same.
(though if you haven't already, it is highly reccomended that your HV should be 120-130vdc)


Well, I am using Steve's transformer without any booster for now so it's not quite that high unless I push it with the Variac.

However, I am a bit confused on this issue. Using your numbers, the load resistance changes by a factor of two and I thought we wanted to keep it at a specific and low value.

Using 5 ohms and 100W as constants, I came up 22V and 4.5A and this is what I am now using although it's more like 5.5A in the real world.

So, using 5 Ohms and 200W I get 31V at 6.3 A. So I am losing 9V of head room.

I seem to be missing something here.

js
K9ACT
Class E Enthusiast
 
Posts: 146
Joined: Sun Jun 27, 2010 12:03 am

Re: Power Supply

Postby kf1z » Thu Sep 09, 2010 6:40 am

K9ACT wrote:However, I am a bit confused on this issue. Using your numbers, the load resistance changes by a factor of two and I thought we wanted to keep it at a specific and low value.


Yes, forgot you are using a PWM... with class-H this wouldn't be an issue.

K9ACT wrote:
Using 5 ohms and 100W as constants, I came up 22V and 4.5A and this is what I am now using although it's more like 5.5A in the real world.


You mean to actually have 100 watts output from the transmitter I assume.
Because we are talking about INPUT power here, as far as the PWM filter is concerned.

K9ACT wrote:So, using 5 Ohms and 200W I get 31V at 6.3 A. So I am losing 9V of head room.

I seem to be missing something here.

js



No, you're not missing anything.

You are correct, sorry I forgot the PWM filter.......

So the "easiest " thing for you to do, is add a 'boost' transformer, as Steve calls it.

So that you can keep your 5 ohm target, and all your headroom.

Other than that, you would need to change your PWM filter to accomodate a different impedance.

You should however, still be able to hit more than 150% modulation with that transformer you have...
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Re: Power Supply

Postby steve_qix » Thu Sep 09, 2010 6:52 am

Hi Jack,

I'll check out that search code, although a Google search on "shunt capacitor" and "class E" did bring up my site... but, that being said, a better search would definitely be an improvement!!

The shunt capacitor (sometimes called C1 in many diagrams) is a class E thing, although there may be an analog in class D - would have to check the schematic. Anyway, in class E, the shunt capacitor is part of the output network, and it is "shunted" across the MOSFET drains to ground (source). Its primary function is to keep the drain voltage down, and spread out the class E peaks over a greater portion of the RF cycle. We are generally using multilayer ceramic capacitors, which can handle 12 amperes of RF current (much more than a doorknob!). And, they are not expensive - $5 to $7 each or thereabouts. The shunt capacitor is a non-critical value, and a wide range of values will usually work (within 50% of what is ideal).

I bet things are working a whole lot better with that scope probe ground removed from the floating common of the modulator - the output ! Think of the current that must have been flowing in that circuit on peaks !!!!!

On the power supply / transformer /power increase... The transformer, without the boost should be able to deliver between 110 and 120 VDC output (at least that's what I get) depending on the line voltage. So, you *should* be able to run 40 volts DC at carrier coming from the modulator and have plenty of headroom for at least 175% positive peaks.

If the filter is designed for 5 ohms, in theory with 40 volts that's 8 amperes (320 watts input). I don't know if the filter will take the current or not... depends on the core materials in use. If there are 8 MOSFETs in the RF amplifier, 320 watts in is fine. You would have to add another modulator MOSFET and damper diode. The absolute best way to set the power in a PWM transmitter is to vary the input voltage to the power supply. Doing this maintains the carrier to positive peak ratio, and keeps the load on the PWM filter constant. Then it's easy: 35v @7A, 30V @ 6A, etc..

Regards,

Steve




Regards,

Steve
Visit the class E web site at:

http://www.classeradio.com
User avatar
steve_qix
Class E Guru
 
Posts: 434
Joined: Fri Jan 31, 2003 11:03 am
Location: Townsend MA

Re: Power Supply

Postby kf1z » Thu Sep 09, 2010 6:54 am

K9ACT wrote:
steve_qix wrote:
Only people who had implementation problems or bad components (usually bad shunt capacitors - doorknobs are BAD in this application) had problems.


Glad you mentioned that as I am getting by with ceramics but might want to use something else for a bigger version and doorknobs come to mind.

This also brings up another problem I have. I don't know what a shunt capacitor is but I know I have seen it on your web site but I can not find it.
js



Shunt capacitors are used in CLASS-E amplifiers.
They "shunt" voltage between the Drain and source of the RF mosfets.
The higher the capacitance, the lower the peak voltage that appears on the drains.


They are NOT used in Class-D. Which is what you are using...correct?
User avatar
kf1z
Senior Contributer
 
Posts: 558
Joined: Sat Mar 18, 2006 7:07 am
Location: Ely, NV.

Previous

Return to General Chit-Chat

Who is online

Users browsing this forum: No registered users and 1 guest

cron