Expensive diode limiter?

[alk509]

Don't get me wrong, I understand how capitalism works... I'm not complaining or saying that people shouldn't buy the things, I'm just amazed that people are giving MercAudio $225, and getting 8 zeners back.

It's like this: I've heard that a pack of cigarettes costs a couple hundred bucks in prison, so if you smoke and you're in prison, you pay up. Whether or not the smuggler's discomfort from shoving a pack of Camels up his ass is worth exactly one hundred and ninety five dollars and fifty one cents is irrelevant to the fact that I, as an outsider, am amazed that people pay 200 bucks for a $4.49 pack of cigarettes.

Peace,
Al.

 

[Sleeper]

I hear you Al. I don't wan't to have to shove a pack of zeners up my ass.
I'm occasionally amazed that I'll get 200 bones to fix a loose wire or at times just change a few light bulbs.

I do want to make a few of these limiters though.
Any tips?

Cheers
Sleeper.

 

[chrissugar]

[quote author="Sleeper"]
I'm occasionally amazed that I'll get 200 bones to fix a loose wire or at times just change a few light bulbs.
[/quote]

Wow, 200$. I think I will move there and leave the whole music business. I'm very good at changing light bulbs and fixing wires. :grin: :wink:

chrissugar

 

[Sleeper]

Did I mention that sometimes the 200 is for crawling around on your hands and knees in dirty dusty attics so f**kn hot you'll literally be soaking wet within 15 minutes OR it's for doing the 50 meter belly crawl under a house with a mere 12" crawlspace, spiders, and ratshit. At these moments, I'm not at all surprised someone forked out 200 dollahs.

Sleeper.

p.s. can somone check me out here?

Assuming there's a back to back pair of zeners similar to that seen in the green pre input section in the tube.

http://1176neve.tripod.com/id10.html

Although I've built a green pre and seen this configuration before, I don't understand how it works quite well enough to make adjustments for a specific input limitation. I assume you use zeners with a rating that equals the voltage equivalent of the db you're after.

so 18dbv has a peak to peak voltage around 22.
http://www.sengpielaudio.com/calculator-db-volt.htm

so 4 22volt zeners, 1 back to back pair off of hot, 1 off of cold, and tied at ground would do it?

or an alternative?

What about a limiter with some germaniums.
how about figure1 from this article minus the input and output pots.
http://headwize.com/projects/showfile.php?file=limiter_prj.htm

I'm missing something here though. (it seems some of the information in that article is clipped Wink )

I need to test forward bias voltage and then what?
looking at figure 1 what would be the formula to determine r1 if I wanted, for example, 18dbv

also, can you stack germaniums in series like you do with zeners to increase the forward bias voltage? or does the stack just smooth out the clipping.

And then to make this a balanced circuit, It should be component matched and doubled up on both hot and cold lines right?

 

[NewYorkDave]

First of all, if we're trying to emulate the Prism gadget, it's supposed to start clamping at 18dBU, not 18dBV. 18dBU is 17.4 volts peak-peak. Since each zener in the back-to-back pair clamps one half of the waveform, we're interested in the PEAK voltage, which is 8.7V. Subtract from this the FORWARD drop (0.6V) of the other zener diode that's in series with the one that's "zenering" at any given signal peak and that leaves 8.1V. 8.2V seems to be the closest standard zener value.

The clamp should go between hot and cold, not between hot and ground and cold and ground. For reference, we're talking about the sort of circuit shown in figure 5A of that Headwize link. For balanced use, you'd split the value of the current-limiting resistor R1 and place half in series with hot and half in series with cold. As for the value of R1, that would be chosen to pass the desired current through the zeners at the point they begin clamping (thus affecting the "knee" characteristic) without presenting an excessive load to the driving device. It's not a good idea to run a clamp like this without current-limiting resistance since when it's limiting, it's essentially shorting out the output of the device driving it :?

As an aside, I wonder about their choice of 18dBU since a system should usually be set up with 20dB of headroom above nominal operating level, which would mean a clip point of greater than +24dBU in a "+4" system. Perhaps they made the knee so soft that they needed to set the initial limiting point several decibels lower than that to ensure effective suppression of +24dBU peaks.

 

[bcarso]

Sounds as if the device would interact with the source impedance a lot, depending.

Now what would be cooler would be one with an adjustment for clip level, and an indicator (obvious candidate a very efficient LED). Put the LED in a bridge in series with the selectable zener. Or use a transistor Vbe multiplier and a potentiometer. At once you have a much better inherent match since you are using only one zener or other clamp circuit, and probably well-matched diodes in the bridge. The LED and the diodes would soften the knee a bit more as well.

You set the clip level knowing what your system headroom is, and then see how often you are hitting clipping by looking at the LED.

Any interested DIY'ers to try this?

Brad

Note added in edit mode: On reflection the Vbe multiplier is not as suitable as the divider current is substantial. If the divider can be made high Z it might get to be adequate.

 

[Gus]

I have built fuzzes with a bridge and different diodes in the middle like posted just above. The thing is it still fuzzed the signal when it clipped Si diodes or leds with Ges your heard more of the note with the fuzz on top. It also fuzzed the signal in a more unmusical way the greater clean level to clip sounded bad.

I am still tring to understand how a diode clipper can sound good. I would guess it has a resistor(s) to limit the current like Dave posted and adding a resistor in series with a hard clipping device can soften the "edges"

Does this thing even sound good?

 

[bcarso]

It's going to be pretty brutal I suspect no matter what, unless the knee is contrived of some sort of slope-breakpoint approximation. And then that arrangement has to cut in earlier so that it is truly limiting at higher levels effectively, so you are incurring a lot of excess distortion earlier on.

 

[bcarso]

BTW...I did some sims of the LED/Vbe multiplier approach and it looks like there is a slight advantage in terms of reduction of higher-order harmonics compared to the zener approach, but nothing dramatic. It was helpful to use a darlington as the transistor so the resistor divider could be much higher Z and not load the signal all the time.

But the arrangement could be adjustable which is interesting I think.

Another thought occurs: make a signal-powered system that does a real compressor-limiter-clipper. There's a tough engineering task, if it is not to load the source too much and create distortion below limiting/clipping. One could use the techniques from power factor correction now common for mains supplies to make the load look resistive, in order to incur a bit of attenuation while stealing the power to run the thing.

Madness of course, but I had to share it. Now that little assembly would be worth some serious money.

But on the other subject, a few thoughts:

All in all, the debate over ripoff vs. value is something difficult to understand until you have been in business for yourself and seen the real cost of mantaining a product line. I remember when I was arguing with my brother years ago about raising the prices on some accessories in the company's lineup. He accused me of being "greedy". We were losing money at the time and eventually went broke owing all and sundry. The main difficulty was the "look" of the products was antiquated and didn't fit into the locations intended with any sort of style. They worked better and were easier to use than the competition, but they would be vetoed most of the time by the designers of the installations.

The products I was raising the price on the most were relatively cheap and involved some clever RF engineering which the competition didn't begin to attempt. So the customers would buy these from us and spend all the rest of the money on the competitors' products. My brother insisted that raising the prices would reduce sales. In fact I don't believe it changed the demand whatsoever.

Normally, there is an optimal price for an item where profit is maximized by being at the best point on the margin vs. sales curve. There is also the "lemon" effect that must be considered, where if something is too cheap no one believes it can be any good.

But there are niche products that few consider worth the effort to duplicate or offer alternatives to---the volumes at best are so low, and a medium-sized company just can't afford the diversion of resources---the "opportunity cost" is too high. JBL, for example, would never make a barrel with a couple of zeners inside, no matter what the margin, other than maybe as a promotional gimmick.

 

[Sleeper]

Hey NYDave,
thanks alot for your input and a clear explanation of I've got a baseline place to start trying a few different ideas. I just couldn't get my head around the thing the way it was explained in the headwize article, so once again big ups :thumb:

[quote author="Gus"]

Does this thing even sound good?[/quote]

Gus, I don't think this is supposed to sound good so much as hang on to something salvagable.

Imagine you're on location recording a meeting between Karl Rove and some Haliburton executives. They are whispering so you have your gain set pretty high. Dick Cheney walks in sits right near your mic and has a coughing fit- smoke filled room and all- total overload. Anyhow, you DO NOT want digital drop outs on this conversation. You can deal with some trashy distortion, it's a pretty ugly conversation to begin with, but I want to know every word that these guys have to say to one another.:twisted:

Sleeper

 

[hodad]

On the technical side, I really dug the limiter article at headwize. I tink I'm going to have a go at making a germanium diode soft limiter to run synths through--might be a good thing to have in case my son (age 2) gets hold of the regen knob on the modular delay again.

On the biz side: it's fine for PS to charge whatever they see fit. And if the market is truly tiny for these things, then the price is probably not that high. If there does turn out to be a market, I think you'll see competition & much lower prices soon (sort of similar to the chnges in the Chinese condenser market--prices dropped noticeably once there were more than one or two players in the business.)

Also, whiners are an important part of the free market economy too. As, I suppose, are those who whine about the whiners.

Tom

 

[bcarso]

Hehe. I see an infinite regress of whiners happening here.

Reminds me of a conversation I had at a metaphysical workshop once, where the people were highly sensitized to anything resembling victimhood/martyrhood. When I mentioned you couldn't buy anything without getting some (undesirable) little croissants at the table outside the meeting room, someone immediately said "Do you want some cheese with that whine?" I thought that a bit of an overreaction...

As far as the best strategy for a low-harmonic instantaneous nonlinear network for this app., I wonder how well just lifting the network from a sine-wave shaper in a function generator would work. I guess probably not that well, as the shaper is optimized for a specific triangle wave amplitude.

Hmmm. Interesting problem in signal theory.

 

[Mendelt]

I can't really add anything to the ripoff vs. business discussion.

But wouldnt something like this work better if you use two zener pairs. One lower voltage pair with a resistor in series for a softer knee paralel with one higher-voltage pair without a resitor to catch the peaks that the first pair lets through?

What are the specs for such a device? What is the maximum signal voltage we should alow and what is the voltage where the first knee should start limiting?

 

[bcarso]

"What are the specs for such a device?"

That's the question---what would we want it to do?

 

[daArry]

Interesting stuff chaps. I was planning some A/D and think this could be cool option to have on a switchable frontend...

I tink I'm going to have a go at making a germanium diode soft limiter to run synths through

Nice. I have some old raythorn t59247's - could these be used?

:thumb:

 

[Sleeper]

[quote author="Mendelt"]
But wouldnt something like this work better if you use two zener pairs. One lower voltage pair with a resistor in series for a softer knee paralel with one higher-voltage pair without a resitor to catch the peaks that the first pair lets through?[/quote]

Good idea, look at headwize article, read NYDaves post on how to do it balanced and you're started.
I was thinking something like a germanium stage followed by a zener stage.

Your voltage limiting point is dependent on your convertors. Set yours to clamp just under AtoD overload level... my RME's for example can be jumpered for different input levels.
Sleeper

 

[gyraf]

Set yours to clamp just under AtoD overload level..

..it would be easier just to have input- and output trim pots..

 

[Mendelt]

[quote author="gyraf"]
..it would be easier just to have input- and output trim pots..[/quote]

If you're going for the zener with resistor setup trimming the resistors would just result in setting the abruptness of the knee i think.
You could atenuate the signal before and after the zeners but wouldn't you get impedance issues with a passive volume control here? I think it would be easier and cheaper to build one device for each clipping-level you need.

 

[gyraf]

or banks of switched zeners..

..I don't think potentiometer in and out will affect limiting in any serious way - but it's really not hard to throw together a proto and do some tests..

 

[bcarso]

As I mentioned the Vbe multiplier sitting in a bridge looks quite similar to a zener, and is readily and continuously adjustable, as long as you don't want to go too low in voltage. Symmetry is assured if either the zener or Vbe mult is in a bridge.

Putting an LED in the collector as an indicator is just frosting on the cake.