A different "single cycle" problem - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 20th May 2007, 04:49 PM   #1
diyAudio Member
 
Join Date: Mar 2007
Location: Canandaigua, NY USA
Default A different "single cycle" problem

I don't want to hijack the output coil threads, so am starting a new one. Though I simulate using SwitchercadIII, I'm more of a "test reality" soul. In playing with different values and constructions of output networks over the weekend, I found that there are varying degrees of power amp stability. What looks great on the scope, and even measures ok, can still have some very low level HF oscillation going on. Worse, this is often amplitude and load dependent, and often occurs asymmetrically. 5MHz seems to be popular on my bench. It's seems like it may be provoked with complex loads and non-sinusoidal signals more easily than simpler bench tests, but it's there and gone too quickly to analyze. What I really want to know is what's happening in my living room, with all the components lashed together, not on my test bench. No doubt this is something that could be found by back driving the amp with a sig gen, but I wanted to float another idea.

I want to build a radio. Well, not quite, but how about a small box with a multipole highpass filter, followed by some gain, a detector, a pulse stretcher, and maybe an output connector. The idea is you can attach this to the amp output and have it immune to anything in the audio band. Should even the slightest signal be generated between 100kHz and, say 20MHz, the LED would light and stay lit for a few seconds. You could also put a scope on the output and correlate the RF bursts with the incoming audio signal. Ideally, you'd see nothing, but I've had more than one amp where I think this would be useful.

Am I missing the boat here, with some simple test I should be doing but aren't, or is this something that would be useful?

Regards,
Conrad
  Reply With Quote
Old 20th May 2007, 05:39 PM   #2
diyAudio Member
 
destroyer X's Avatar
 
Join Date: Feb 2004
Location: Recife - Brasil Northeast
Default Good idea...but do not forget to shield your amplifier

As Radio frequencies can be picked up into component leads,

Also can enter... travelling over the audio source.

Depending the strength... magnetic waves can be picked into the mains, travelling to the supply and pass over the high inductance electrolitic condensers to have some "fun" into your audio circuit.

Sometimes they are captured over the coaxial shield when you do not have good grounding.

Even your body is a radio frequency receiver... as you use to capture the magnetic lines from the mains.... when you touch the amplifier input with your fingers...you are injecting the 60 hertz ro 50 hertz from the mains, you have captured, into the amplifier input.

Radio frequency is alike US missiles...they can enter the door hole you use to put your key to open and luck your door....you can put traps everywhere...they will enter the back door.... i am a Radio Amateur, since 1965 on the air...maybe you too..and we know the hell those things are.

Depending the frequency...can enter 2 mm diameter holes.

A faraday cage...bad grounded is a wonderfull receiver...nice antenna it can be..depending how good or how bad the ground is.

Well....of course you know that...but it is good to remember.

In your place, Conrad....if you construct a big coil and a rectifier..you may drive a small motor...just capturing the magnetic wave you have with strong intensity in your place..one of the most Radio Frequency poluted places in earth...the detected energy may move those micro motors that need 3 miliamps and 300 milivolts.

regards,

Carlos
__________________
These words sounds alike English..but they are not.. these are words without meaning, just sounds made by humans; https://www.youtube.com/watch?v=_X_7iMHugXM
  Reply With Quote
Old 20th May 2007, 06:09 PM   #3
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Default Re: A different "single cycle" problem

Quote:
Originally posted by Conrad Hoffman
I don't want to hijack the output coil threads, so am starting a new one. Though I simulate using SwitchercadIII, I'm more of a "test reality" soul. In playing with different values and constructions of output networks over the weekend, I found that there are varying degrees of power amp stability. What looks great on the scope, and even measures ok, can still have some very low level HF oscillation going on. Worse, this is often amplitude and load dependent, and often occurs asymmetrically. 5MHz seems to be popular on my bench. It's seems like it may be provoked with complex loads and non-sinusoidal signals more easily than simpler bench tests, but it's there and gone too quickly to analyze. What I really want to know is what's happening in my living room, with all the components lashed together, not on my test bench. No doubt this is something that could be found by back driving the amp with a sig gen, but I wanted to float another idea.

I want to build a radio. Well, not quite, but how about a small box with a multipole highpass filter, followed by some gain, a detector, a pulse stretcher, and maybe an output connector. The idea is you can attach this to the amp output and have it immune to anything in the audio band. Should even the slightest signal be generated between 100kHz and, say 20MHz, the LED would light and stay lit for a few seconds. You could also put a scope on the output and correlate the RF bursts with the incoming audio signal. Ideally, you'd see nothing, but I've had more than one amp where I think this would be useful.

Am I missing the boat here, with some simple test I should be doing but aren't, or is this something that would be useful?

Regards,
Conrad
Conrad,

Interesting idea. To get an idea how it looks before having to commit time and money, could you loan somewhere a ham receiver? These things often can receive frequencies from the 100's of kHz to many 10's of MHz. Connect the speaker output of your amp to the antenne input of the receiver by a small capacitor and listen if you 'receive' anything.
One proble might be to screen the receiver from real rf, possibly you need to screen the antenna input and connect it with a well screened cable to the amp.
Just thinking out loud here.

Jan Didden
__________________
If you don't change your beliefs, your life will be like this forever. Is that good news? - W S Maugham
Check out Linear Audio!
  Reply With Quote
Old 20th May 2007, 07:13 PM   #4
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Sorry to just barge in like this. I was doing a search for sources or construction methods for output inductors/coils, and happened-across your thread.

You probably already know all of this, but...

Perhaps you do have some high-frequency instability. But, with LTspice (aka SwitcherCad), I can usually actually see it, if I do square-wave tests, and have a complex speaker model for the load instead of just a resistance (but often also with just a resistance), and maybe also an RLC model for the speaker cables. (See http://sound.westhost.com/cable-z.htm , for the speaker and cable models, for example.) And, of course, I usually have to magnify the heck out of the corners of the output square wave from a transient run, to really see what's there (although very slow periods of simulation speed are usually a good clue). You also have to set the max timestep quite small, to see high-frequency stuff reliably; probably 10e-6 or lower.

Maybe more-importantly, I usually also try to model as many parasitics as are practical, such as ESR of all caps (which varies with frequency and temp, IIRC; can be a pain; usually I just use spec'd value, at least at first), DCR of inductors, capacitance parallel to each resistor (0.3 to 0.5 pF), and the parasitic R and L in series with traces and wires (the web has calculators and tables for them), and C-to-whatever if there's a ground plane or nearby traces (ditto).

Getting back to your main topic, i.e. some "receiver" on a real output, to detect the presence of very high (relative to audio) frequencies:

There are also some more-or-less off-the-shelf solutions. Maybe you could acquire an older HP 400-series (or other) wideband analog RF Power Meter, and put a high-pass filter in front of it. They are often dirt-cheap, now, and are small and light weight. (Or, download the technical manual for one and see how they designed them. There are lots and lots of RF-related tech manuals at http://bama.sbc.edu/hp.htm .) If you want to try to acquire one, try either http://www.govliquidation.com or http://www.ebay.com , and maybe http://www.labx.com .

There were also units such as the HP 3586C "Selective Level Meter", which were essentially radio receivers. They did more than you currently are asking for, since they were tunable with a very fine resolution (something like 0.1 Hz!). And I think they are still quite expensive. Of course, a spectrum analyzer would be great, too, but also gives much more-detailed information than you're currently asking for, and would also usually be relatively expensive.

I guess the question should also be asked as to whether or not a low-level 5 MHz (for example) oscillation even matters. It guess it might matter to the FCC. And I suppose it might get radiated and interfere with other devices, somewhere. But that would depend on its strength, etc. And it sounds like you're talking about quite low amplitudes. But does it noticably affect the operation of your amp? I guess "it depends". If it's truly an instability, and under some condition it starts oscillating well, something might get very hot, very quickly. Also, high frequencies (e.g. RF) can cause other problems, since they tend to get rectified by semiconductor junctions and can end up causing DC offsets, for example.

It's usually considered good practice to put an RF filter at the input of every amplification circuit (e.g. every opamp). This is covered fairly well in a recent book that Walt Jung was editor of, which is available in its entirety on analog.com . It's called "Op Amp Applications Handbook". But the RFI/EMI principles should apply more-broadly. Chapter 7 has a lot of good material about EMI/RFI considerations. The book also has a lot of good broadly-applicable audio amplifier material.

I suppose it's also possible (probably quite likely, actually) that you are seeing something that's externally produced, and is being received by your circuitry. You might have to build a "Faraday Cage", to be certain of the source. (Sounds like a lot of work...)

Your original idea sounds pretty good, to me, especially if it were easy and cheap to make something that also worked well. As I hinted at, before, you might want to search for designs for "RF Power Meter" circuits, and then just use the ideas for the first stages, if you just want a "yes/no"-type of indicator for the presence of RF, instead of actual measurements, which should make it cheaper and easier. You might want to search at http://groups.google.com (which has the searchable message-traffic archive of Usenet newsgroups, with basically all traffic since about 1981! A goldmine!), and then maybe post some questions in the sci.electronics.design group, and maybe also one of the rec.radio.amateur.* groups. Of course, you should also google the web for something like "RF Power Meter OR Indicator circuit OR schematic OR design" (without the quotes, but with the OR's in caps). Oh, also, when doing both of those google searches, you might want to append "-fs -fa -sale -auction", if you're not looking for ads for units for sale.

I forgot to mention, too, but Jan Didden already has done so, that a ham/amateur radio receiver, or maybe even a generic shortwave radio receiver, might be helpful.

By the way, if you haven't already done so, you will probably want to check out the LT-SPICE discussion group, at yahoogroups.com. It's excellent!

Good luck!

- Tom Gootee

http://www.fullnet.com/~tomg/index.html

-
  Reply With Quote
Old 20th May 2007, 07:35 PM   #5
diyAudio Member
 
Join Date: Mar 2007
Location: Canandaigua, NY USA
A ton of good ideas! I do intend to own one of everything ever made before I expire, and am well on the way. I have a Sony 2010 receiver that goes from a couple hundred kHz to 20Mhz, also an HP selective level meter. No SA, other than my sound card. The problem with all this is the frequency is unknown, and it occurs in short unpredictable bursts. I need a wide band detector that just grabs everything, even short bursts. Thus my desire for something with a bit of hold- IMO, it's too easy to miss with SA. SA is great for steady state stuff when you can do a lot of averaging, but pretty bad for transient problems. My guess is I can throw something together pretty fast that will do the job. Maybe the easiest thing to do really is provoke the instability on the bench so it can be analyzed steady state. I've done that to some extent, but now I'm having a crisis of confidence that the monster has really been beaten back into the swamp, never to return. The amp I'm fooling with is the old SWTP Tiger design, and IMO it's a good reliable performer, but has never been as free of HF instability as I'd like. I built it back in my younger days, in a nice tidy package that was never stable and impossible to work on. I tore it all apart and rebuilt it into a huge rack chassis, so it makes a great development mule for any idea I want to try.

Gootee, you're a man after my own heart. I've made several posts about the futility of simulations that don't include the parasitics, and it sounds like you're way ahead of most. Most people seem to have no grasp of wire inductance and the imperfections of real world components. Hijacking my own thread, here's a question no one has been able to answer for me- When you model a cap, the dissipation factor is more or less constant with frequency. The ESR is not. LTSpice only lets you enter the ESR, so IMO frequency sweeps are missing an important variable. How do you take the real ESR into account? It seems you need a frequency dependent resistor, or is there some built in method I haven't spotted?
  Reply With Quote
Old 20th May 2007, 07:40 PM   #6
diyAudio Member
 
jan.didden's Avatar
 
Join Date: May 2002
Location: Great City of Turnhout, Belgium
Blog Entries: 7
Quote:
Originally posted by gootee
[snip]I suppose it's also possible (probably quite likely, actually) that you are seeing something that's externally produced, and is being received by your circuitry. [snip]- Tom Gootee

http://www.fullnet.com/~tomg/index.html

-

I've had cases where such of type a low level hf oscillation actually came from the 'scope itself...

Jan Didden
__________________
If you don't change your beliefs, your life will be like this forever. Is that good news? - W S Maugham
Check out Linear Audio!
  Reply With Quote
Old 20th May 2007, 07:43 PM   #7
Hartono is offline Hartono  Indonesia
diyAudio Member
 
Join Date: Nov 2002
Location: currently in China
I think any textbook already explained this.

the amp subjected to such signal will go to slew limit, the feedback will not be fast enough to do it's job , and on that period of time when the amp slew limit ,the amp will essentially be open loop.

It might not exactly be called instability, it just distort.

Hartono
  Reply With Quote
Old 20th May 2007, 07:54 PM   #8
Hartono is offline Hartono  Indonesia
diyAudio Member
 
Join Date: Nov 2002
Location: currently in China
about the varying ESR with respect to frequency

we can model with some degree of success by adding ESL (inductor)


Hartono
  Reply With Quote
Old 20th May 2007, 08:11 PM   #9
diyAudio Member
 
Join Date: Sep 2006
Default Re: A different "single cycle" problem

Quote:
Originally posted by Conrad Hoffman
I don't want to hijack the output coil threads, so am starting a new one. Though I simulate using SwitchercadIII, I'm more of a "test reality" soul. In playing with different values and constructions of output networks over the weekend, I found that there are varying degrees of power amp stability. What looks great on the scope, and even measures ok, can still have some very low level HF oscillation going on. Worse, this is often amplitude and load dependent, and often occurs asymmetrically. 5MHz seems to be popular on my bench. It's seems like it may be provoked with complex loads and non-sinusoidal signals more easily than simpler bench tests, but it's there and gone too quickly to analyze. What I really want to know is what's happening in my living room, with all the components lashed together, not on my test bench. No doubt this is something that could be found by back driving the amp with a sig gen, but I wanted to float another idea.

I want to build a radio. Well, not quite, but how about a small box with a multipole highpass filter, followed by some gain, a detector, a pulse stretcher, and maybe an output connector. The idea is you can attach this to the amp output and have it immune to anything in the audio band. Should even the slightest signal be generated between 100kHz and, say 20MHz, the LED would light and stay lit for a few seconds. You could also put a scope on the output and correlate the RF bursts with the incoming audio signal. Ideally, you'd see nothing, but I've had more than one amp where I think this would be useful.

Am I missing the boat here, with some simple test I should be doing but aren't, or is this something that would be useful?

Regards,
Conrad

You're not missing the boat at all.

First, I have seen such low-level oscillations as you have.

Second, I actually started to build the "parasitic oscillation sniffer" that you pretty much describe a couple months back, but did not yet finish it.

Great minds think alike!

Bob
  Reply With Quote
Old 20th May 2007, 08:56 PM   #10
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally posted by Conrad Hoffman
A ton of good ideas! I do intend to own one of everything ever made before I expire, and am well on the way. I have a Sony 2010 receiver that goes from a couple hundred kHz to 20Mhz, also an HP selective level meter. No SA, other than my sound card. The problem with all this is the frequency is unknown, and it occurs in short unpredictable bursts. I need a wide band detector that just grabs everything, even short bursts. Thus my desire for something with a bit of hold- IMO, it's too easy to miss with SA. SA is great for steady state stuff when you can do a lot of averaging, but pretty bad for transient problems. My guess is I can throw something together pretty fast that will do the job. Maybe the easiest thing to do really is provoke the instability on the bench so it can be analyzed steady state. I've done that to some extent, but now I'm having a crisis of confidence that the monster has really been beaten back into the swamp, never to return. The amp I'm fooling with is the old SWTP Tiger design, and IMO it's a good reliable performer, but has never been as free of HF instability as I'd like. I built it back in my younger days, in a nice tidy package that was never stable and impossible to work on. I tore it all apart and rebuilt it into a huge rack chassis, so it makes a great development mule for any idea I want to try.
Back when I used to buy and resell mil surplus test equipment (talk about a kid with his own candy store!), I got to play with some Tektronix 7L12 or 7L13 (and other) spectrum analyzers, for short periods of time. The 7L12 or 7L13 has some sort of "peak hold" mode, where the highest level yet-recorded is continuously displayed, for each frequency. That might work, although I don't know how much signal energy was required to even "register", for that. I only remember doing it with an FM antenna connected to the input, so I could see all of the FM stations' peak levels, all at once. :-)

Quote:
Originally posted by Conrad Hoffman

Gootee, you're a man after my own heart. I've made several posts about the futility of simulations that don't include the parasitics, and it sounds like you're way ahead of most. Most people seem to have no grasp of wire inductance and the imperfections of real world components. Hijacking my own thread, here's a question no one has been able to answer for me- When you model a cap, the dissipation factor is more or less constant with frequency. The ESR is not. LTSpice only lets you enter the ESR, so IMO frequency sweeps are missing an important variable. How do you take the real ESR into account? It seems you need a frequency dependent resistor, or is there some built in method I haven't spotted?
[Ahhhhh! I'm basking in your appreciation! Thanks!]

Regarding the ESR vs freq for LT-Spice simulations, I have only done it by using separate simulation runs for different frequencies, with the ESRs set by hand for each capacitor for each simulation run.

However, I never did get serious about finding a better way to do it. It SEEMS like it "should" work to use a variable (e.g. {ESR}) in the ESR field, if you had an equation for ESR vs freq, probably in .param statement(s). I guess that then you'd also need a different ESR variable and equation for each type of capacitor.

That would probably be a VERY good question to ask in the LT-SPICE discussion group, at yahoogroups.com (after first searching the message archive, of course). There are some very brilliant people who frequent that group, who seem to know almost everything about getting the most out of LTspice (and about a lot of other stuff!). The author of LTspice also is there quite often, Mike Englehardt, of Linear Technology Corp, aka "Panama Mike" in the group. The main moderator, there, the extremely-impressive LTspice guru Helmut Sennewald, will probably be able to answer the question more-or-less immediately, if someone doesn't beat him to it.

By the way: Since you're into "equipment", you might also really like the TekScopes group and the hp_agilent_equipment group, both at yahoogroups.com . TekScopes, which I frequented more, had some incredibly-great people on it. (And since it sounds like you still want to ACQUIRE more equipment, you _really_ ought to check out govliquidation.com. I think they might have some of the good warehouse sites near you; ones that have the big test equipment sales, monthly. JUST going to the pre-sale inspection period is AWESOME, at least if they have a couple-hundred auction-lots or more, there. You can used their Advanced Search and specify FSC code 6625, for most test equipment. DC power supplies might be 6130, IIRC. But there's a complete list of the codes, there, too. You should also go to the warehouse-site part of the search and hold down Ctrl and hit only the ones near-enough to you. But be prepared to sink a lot of TIME. It's addictive.)

Regarding parasitics modeling in Spice: I still consider myself to be "a newbie", with spice; just two or three years in, so far, and usually too busy to learn the best ways of doing things, with it, unfortunately. Having read quite a few manufacturers' appnotes that mentioned parasitics, I just happened to have some designs I was working on, with LTspice, a couple of years ago, where I thought parasitics might really matter. And they did, in some cases. So I got much more interested in them. One of the last things I started trying to use the modeling of parasitics for was trying to determine when to add separate power and ground cabling, which had to run from a power supply board to separate circuit boards. Modeling the parasitics of the power and ground wiring, and some of the traces, and then trying various combinations of sharing them, made it pretty obvious when a board or circuit might need more than one set of power and/or ground wiring.

Then I used the same circuit model to try to find out how to lower the levels of disturbances on the power and ground rails, which I thought were probably being caused by certain fast and/or high power circuits that shared the same power supply, and were "disturbing" some of the more-sensitive circuitry. I had tried just about _everything_ else I could think of, without much success. In that case, I "discovered" that putting certain small-ish chokes (sometimes with a small parallel resistance) in series with the power inputs to certain devices or circuits, just before their bypass caps, could make the REST of the power (and ground) rails much, much quieter. I wish I could find a paper or some appnotes about that sort of thing. Or maybe it's obvious to everyone but me? (I was an EE starting in 1980, but in automatic control theory somewhat more than circuits, for a few years, but then spent 15 or more years writing non-EE software, and have only been dabbling in electronic circuits, again, for the last few years.)

Sorry to have blathered-on about all of that, for so long.

[/hijack] :-)

- Tom Gootee

http://www/fullnet.com/~tomg/index.html

-
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
"Mighty Mite" Single Ended Amplifier wrenchone Tubes / Valves 13 20th October 2008 04:57 AM
glass audio 1995 "Single Ended Glory" habsrock93 Tubes / Valves 3 22nd June 2008 02:10 AM
FS: Homebuilt "Aikido" 2A3 single ended triode amp jim_vt Swap Meet 11 21st January 2006 03:38 PM
"lift off " or "Houston we have a problem" tenderland Tubes / Valves 16 1st May 2005 05:08 PM
Using Telefunken "Tank" transfomers for a single ended el84 amp? ark Tubes / Valves 13 8th March 2004 11:32 PM


New To Site? Need Help?

All times are GMT. The time now is 12:05 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2