www.wavebourn.com

We create creativity!
It is currently Wed Jun 26, 2019 9:45 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 187 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13  Next
Author Message
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 5:28 am 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
wavebourn wrote:
Baking tubes in an oven worth it: instead of absorption of gases by cathodes it is better to absorb them by getter flash.

The way it was explained to me by my mentors was that if a tube has gettering designed in as part of the cathode processes, then you should run the cathode to 'reactivate' the tube. This is mostly the case for very large tubes with oxide cathodes.

And for those, who doesn't have this functionality, then it won't do any harm. Instead you just use the heaters to heat the flash gettering to a suitably high temperature, while simultaneously reactivating the emission properties of the cathode to full specs (if possible).

Would very much like to know if I have misunderstood something. The oven trick should work as well, at least for tubes without bakelite sockets, like most Octal tubes have. What time and temperature would you recommend? Can I bake a pizza at the same time? Cool

quikie22 wrote:
Actually it was a combination of two things... newbie + too excited to get the thing up and running...

My previous experience with the smaller tubes were pretty much ok, hence my confidence in hooking up the GU-50 immediately.... unfortunately it didn't work out as planned.

I suspect you are not the first one, who jumped to that conclusion. Very Happy

Unfortunately it seems some people just assume tubes are either OK out of the box, or they are permanently dead. I do wonder now and again how many salvageable tubes have been discarded as trash in this way over time...? Sad


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 5:58 am 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
Decided to test using an electronic Gyrator in place of the PSU choke. This was a case, where Spice really did the trick, as the simplicity of the schematic belittles the effort I had to spend to get this 'just right'. However the final performance is pretty nice, and completely blows the expensive and heavy choke out of the water. With a 4500 ohm DC load in parallel with 100mApeak at 10Hz, the ripple on the regulated output is 30mVrms. Without the signal component of the load the mains ripple is below 1mVpeak-peak. Suspect physical layout will be the limiting factor for actual real world performance.

Plot below is worst case scenario when starting up, hot heaters yet with zero volts on the capacitors. Power dissipated in MOSFET peaks at 22W yet quickly settles at a steady 3W. Output ripple isn't visible on the green trace at this scale. The noise is just rounding errors in the plotting routine.

Image

Image


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 6:26 am 
Offline

Joined: Mon Mar 22, 2010 1:10 am
Posts: 238
Why do you use a gyrator in PSU? Gyrator would actually cause power to sag on sudden current draw on sudden peaks of AB2.

I thought it is a source follower that we should use to provide a low output impedance with good ripple reduction and possibly voltage regulation. Correct me if I am wrong.

Somehow in real life testing of both the voltage regulator source follower against the choke, the choke offers a nicer sound Smile However can't comment on the gyrator as have not tried it in PSU before.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 7:02 am 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
quikie22 wrote:
Why do you use a gyrator in PSU? Gyrator would actually cause power to sag on sudden current draw on sudden peaks of AB2.

Two things spring to mind.

1) Yes, it would sag during prolonged current peaks when running class AB(2). However my planned output stage should run 100% class A, where any significant change in current would be due to excessive even order harmonics. Also, the grid current in my design, the '2' part, is supplied by my separate +/- 30V rails.

2) A properly tweaked filter using a genuine choke would sag nearly as much as this design, if it was used for an amplifier with uneven current load. If you look at my previous schematic for my PSU, there is a 58 ohm resistor in series with the 5H choke, the choke's loss resistance. If I actually used that filter with a changing load current, the output voltage would 'ring' long after the load change had stopped. This is because the choke is said to be under damped as I used it.

To be critically damped, which results in minimum ringing, you need a total series resistance of

Rcrit = sqrt(L/C)

where L is the inductance of the choke in Henries, and C is the capacitance in Farad at the output of the filter. Here Rcrit = sqrt(5H/110E-6F) = 213 ohm. So I am still missing about 150 ohm in series with the choke. For an average load current of 100mA, I'd loose about 22 V across the choke, compared to about 30 for the gyrator. Nearly identical, in other words, if we want to minimize ringing when changing load current.

quikie22 wrote:
I thought it is a source follower that we should use to provide a low output impedance with good ripple reduction and possibly voltage regulation. Correct me if I am wrong.

Look closely. They are the same thing here. I just use a filtered yet floating voltage reference by way of C5 and C6 for the source follower.

quikie22 wrote:
Somehow in real life testing of both the voltage regulator source follower against the choke, the choke offers a nicer sound Smile However can't comment on the gyrator as have not tried it in PSU before.

Nor have I for audio amps, I am just tinkering here. No idea really how any of this will sound. Very Happy However my gyrator design would create problems beyond a simple voltage sag if the load current changed drastically, say going temporarily from 100 to 150mA and then back down after some time. Particularly on the 'downswing' some nasty problems might raise their heads, which I suspect might lead to unwanted audible effects.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 10:31 am 
Offline
Шаман, типа...

Joined: Tue Feb 10, 2004 12:30 pm
Posts: 7321
Location: Pleasant Hill, California
It is not a gyrator. It is just a source follower for the filtered voltage. No harm done. Gyrator would have a positive feedback through capacitor.

_________________
А женщина даже в мужской рубашке выглядит гораздо приличнее, чем мужчина в женских колготках! Be-be-be!


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 12:10 pm 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
My bad, thanks. Embarassed

I just hope the floating source follower will work as well in reality as it does in the simulation. Would be nice to be able to drop the pair of chokes from the project.

Probably not the last mistake I'll make either. Right now I am sitting here, poking MOSFET datasheets and scratching my head. I was trying to decide which high voltage device to settle on for all my needs in this project. But I have hit a little snag of sorts. If my reading skills are not completely hopeless, then - for low currents - a low current device is preferable for a source follower configuration compared to a high current MOSFET. On the low current device the gate voltage needs to increase slightly less for a drain/source current increase from 0.1 to 1A. Strange...

Not that I am complaining, mind, as the low current devices costs 3 times less than their bigger brothers.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 9:24 pm 
Offline

Joined: Mon Mar 22, 2010 1:10 am
Posts: 238
A low current MOSFET will have lower Miller capacitance as well IIRC.

However we still need to consider the dissipation of the device. I usually try not to exceed 30% of the max dissipation of the device to ensure stability and longevity.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Thu Mar 10, 2011 11:46 pm 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
Well, I can tell what criteria I was focusing on when choosing the final MOSFET to buy, and how I got to those.

*) Needed to be a fully insulated TO-220 case, as I refuse to use anything else when the device floats at 500V above the heatsink/chassis. This is not the place for mica and grease IMHO. The Insulated TO-220 cases are good for about 2.5KV between device and heatsink right out of the box.

*) Needed to have an accurate Spice model available. The FET model I used above was just a generic one I had in the library for a high voltage device. It is likely to be quite inaccurate. I wanted an accurate model from the device manufacturer themselves, as that is the only way of accurately judging the performance of a source follower without actually buying the device and test it. The plan is to buy a small pile of a single type for all my high voltage needs in the pair of amps. Low voltage regulation would need something else, of course.

*) Must survive when the amp is switched on. Worst case startup of the regulator happens so slowly that we have to use the SOA specifications for DC operation. Peak power during start happens at about Vds = 400V, Ids = 60mA, with maximum voltage just shy of 600V. To reliably survive that, it seems from the SOA curves I have seen we need at least an 800V / 40W device, much beyond what the idle specs of the regulator might suggest.

*) Preferably cheap and readily available to anyone, anywhere, in case somebody else might want to replicate the amp some day.

All said and done I could only find one family of devices, which fulfilled all these criteria, the Fairchild FQPFxN80C series. Fairchild was happy to mail even a little hobbyist like myself all the Spice models I wanted, and on the other hand some of the other manufacturers doesn't even specify SOA for DC operation of their power MOSFETS?!

With the accurate models in hand I checked for the phenomenon mentioned, and indeed. The 3A member of the family gave noticeable better performance in my circuit. The 7A device generated about 2.5 times more load induced ripple. So the final choice fell on the Fairchild FQPF3N80C for them being good citizens with their Spice models, and making a nice device available for $1 in quantities of one. icon_biggrin

With this device the regulator has mains induced ripple down around 0.5E-5 Vpp, which is just silly. For 100mApeak load changes at 10Hz the rail wobbles 100mVrms, equivalent to 0.06%. Miller capacitance as a source follower is roughly 2x Crss, or about 11pF as Crss for this device is 5.5 pF. Cool


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 12:18 am 
Offline

Joined: Mon Mar 22, 2010 1:10 am
Posts: 238
I only have a piece of FQPF part and have yet to try it, although this is only a 600V rated device. I have been trying to find a suitable MOSFET too and have difficulty obtaining them here. Thanks for doing all the groundwork. Smile

Sadly, FQPF components cost quite a bit more than USD1 in my part of the world. The one I have costs USD5, and it was the only one piece that they had Sad


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 12:35 am 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
Where do you live then? If you can wait until I get around to ordering my stash, I could buy some for you as well and mail them. Would be easy to pay for if you have PayPal.

At current exchange rates each would cost $1.21. Shipping shouldn't break the bank, not even if you live in New Zealand. Wink

These prices assume we need at least 25 between the two of us, and I need 8 for both channels even before considering spares and a stash for experimentation/future builds.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 4:37 am 
Offline

Joined: Mon Mar 22, 2010 1:10 am
Posts: 238
Let's just see wht Anatoliy thinks of their specs, cos I am wondering if we should go for the 5A parts? The MOSFETS Anatoliy uses are rated way above what the DC idle parameters are at. Even my voltage regulator made with IRF840 blows once a while even though it is so conservatively rated below it's SOA.

If Anatoliy says so, I wouldn't mind getting a few of these babies. I am actually quite near New Zealand... Malaysia to be precise.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 4:47 am 
Offline

Joined: Mon Mar 22, 2010 1:10 am
Posts: 238
Let's just see wht Anatoliy thinks of their specs, cos I am wondering if we should go for the 5A parts? The MOSFETS Anatoliy uses are rated way above what the DC idle parameters are at. Even my voltage regulator made with IRF840 blows once a while even though it is so conservatively rated below it's SOA.

Also the 1000V parts would probably be better choices too if you plan on getting a bunch of them. You could easily use them for any project up to around 800V easily if you have a desire to do so.

An example of what Anatoliy uses for his high voltage MOSFETs
http://www.datasheetcatalog.org/datashe ... 727-01.pdf

If Anatoliy says so, I wouldn't mind getting a few of these babies (FQPF). I am actually quite near New Zealand... Malaysia to be precise.


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 5:03 am 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
Do you have a link to the exact schematic of the amp you have built, please? Would be interested in seeing how the MOSFETs are protected, if you manage to blow some of them on a regular basis.

And yes, the sneaky plan behind posting the specs and schematics was in the hope that Anatoliy would start making noises if he felt they were insufficient for the job. Wink

The next larger size my supplier has in stock seem to be the FQPF6N80C, 5.5A at 800V. Pd is 51W for the insulated part, and it costs $2.82 each if we buy a total of 25 or more. The cost include Danish sales tax incidentally, which I unfortunately have to collect...

Let us hope Anatoliy doesn't live very close to the Californian coastline. The news from the Pacific are not good today... Crying or Very sad


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 1:44 pm 
Offline
Шаман, типа...

Joined: Tue Feb 10, 2004 12:30 pm
Posts: 7321
Location: Pleasant Hill, California
Knarf wrote:
Let us hope Anatoliy doesn't live very close to the Californian coastline. The news from the Pacific are not good today... Crying or Very sad


Hey, what news? Shock
I live in San Francisco Bay Area. It's quiet and sunny here.

http://www.latimes.com/news/nationworld ... ?track=rss

Speaking of blown MOSFETs, I see 2 major ways: either overcurrent when shorted to ground, or gate blown-up. For the first something like 22 Ohm flame-proof resistor in series with source helps, for the second 12V Zener between gate and source is used. That's it.

_________________
А женщина даже в мужской рубашке выглядит гораздо приличнее, чем мужчина в женских колготках! Be-be-be!


 
 Profile  
 
 Post subject: Re: GU-50 Single Ended Amplifier Design
PostPosted: Fri Mar 11, 2011 2:11 pm 
Offline

Joined: Wed Oct 01, 2008 9:14 am
Posts: 41
Location: Denmark
wavebourn wrote:
Hey, what news? Shock
I live in San Francisco Bay Area. It's quiet and sunny here.

Glad to hear it, the news companies made the situation sound pretty grim for a while.

wavebourn wrote:
Speaking of blown MOSFETs, I see 2 major ways: either overcurrent when shorted to ground, or gate blown-up. For the first something like 22 Ohm flame-proof resistor in series with source helps, for the second 12V Zener between gate and source is used. That's it.

Yep, output shorts could be nasty. I have to use some pretty tough zeners in my circuit for them to survive that. Ordinary 400mW units would just go *pop* in that situation, probably going open circuit and causing a MOSFET failure as a result. I will probably use Vishay BZT03C12-TR or something similar: 600 (six hundred) Watts of non-repetitive pulse energy for 100uS...

Auto-biasing the MOSFET off via a source resistor is a neat trick, except it ruins the regulator output impedance somewhat. Will have to think about that one a bit.


 
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 187 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2002, 2006 phpBB Group
[ Time : 0.045s | 13 Queries | GZIP : Off ]