H11F3 (H11Fx)

[ve3wwg]

Colin: You've mentioned these before.

Today I finally spent some quality time (on my commute home) going over every detail of the Fairchild datasheet. I love everything about this part except for it's price (more on that later).

Obviously (as you said) this part is very linear as the spec sheet claims 99.9% linearity. As a resistor it is capable of about 470 ohms (H11F3) to 300 Megohms as a variable resistor.

The one thing that got me excited about this was Figure 2 Output Characteristics (attached to this note).  Look at those curves!  This is screaming to me-- distortion pedal!!

Back to back diodes have a similar response but the problem with diodes is that the limit is fixed.  What I like about this, is that by varying the drive of the LED you can move those limiting parts of the curves (plus and minus) up and down.

So you could develop an AGC type of voltage based upon the non-distorted guitar signal. This AGC signal could then move the boundary up when the signal amplitude increases. Standard diode distortion just compresses the signal instead.

Now of course, you can limit the amount of level shifting, as there is some advantage to getting more dirt sometimes- but you'd be able to control it!!  I love this idea.

The Holy Grail:

Here is what I really want and I think the H11Fx will give it to me: at the extreme ends of that response curve in Fig 2, I'd like the response to fall closer to zero as the signal gets increasingly too high in amplitude.

Think of a vacuum tube when the grid is driven positive. Yes, the signal squashes a bit when this starts. But if you push the grid extra positive, electrons start being attracted to the grid enough to start stealing current from the plate circuit. In essence, an extra positive grid makes the plate conduction start to degrade somewhat, rather than remain flat.

Why is this good?  It helps to bolster harmonic content because the tip of the sine becomes double humped. I believe this primarily helps the 3rd harmonic, which is the nice stuff (even harmonics come from a non-symmetrical signal).

So rather than a diode that just limits the signal in a flat way, we let the signal fold somewhat, keeping some character present and helping with odd level harmonics.

To do this, I imagine the LDR doing its curve limiting, much like a diode. But you have signal rectifier that only rectifies signals above a small increment (like one diode drop). I.e the tips of the signal only. Then only the tips of the sine wave would be used to subtract LED drive current. This would have the effect of reducing the current limit at the right and left extremes of Fig 2.  In other words- signal folding!!

Hold the Phone!

Ok, there is one important difference (see if you noticed it).  Diodes limit based upon voltage (approx 0.7 volts for silicon). If you look close at Fig 2, it is doing current limiting instead. For example, when the LED drive is 6 mA, the maximum current the receiver will conduct is about +/- 200 uA.  Yet you can see from the bottom axis, the voltage is allowed to extend to +/- 0.2 volts and beyond.

Since you usually want to amplify voltage, you can convert this to a voltage by running the LDR current through a resistor in series. That way when the current maxes out at +/- 200 uA, the voltage across R will max out as well (ohms law).  The slack voltage is developed across the LDR instead.

Price

I see there are some obscure places on the net that seem to offer these at sub $1.00 prices if you buy them in sufficient quantity.  I may end up buying 10-25 of them if I can get them at the right price and shipping.  $2.45 each seems a bit much, though on a single unit basis, this isn't going to break the bank.  But I think I need several of these to mess with.

I need to do signal folding. It's what I want and I've yet to find a clean way to do it. This might be it!

[expanoncolin]

You know, when you asked about the linearity yesterday and I pulled up the datasheet, I saw that exact same figure and thought the same thing - "Wow, looks like a nice distortion curve!"  and the same thought re: current limiting instead of voltage limiting.  I wasn't able to make the mental leap to actually use this in that context.  It certainly sounds like it would make a pretty ideal device for it though.  Looking forward to seeing what you come up with.

In terms of pricing, back in the day I used to be able to get them for about than $1 each, I think from allied electronics, but now it looks like the cheapest name-brand source is Mouser at $2.10 a pop.  Allied still actually has them but they're non-stock:
http://www.alliedelec.com/search/productdetail.aspx?SKU=090SPEC&MPN=H11F3M
The only other place I saw was this no-name site:
http://www.omnipro.net/isocom/H11F3

-Colin

[ve3wwg]

The first one I pulled up at Mouser had a minimum order of 750.

I see a selection criterion of PDIP-6 with Gull Wing.  I've not seen that term before. Sounds like a heat sink but obviously not for this part. Do you know what that is about?

I saw the part at your other link. I found a few other places too.  I may have to try one of them to see if I can get a hobby stock of those parts.

But definitely, I need to get at least one soon to experiment with.  I feel a Mouser order coming on. My wish list is buldging there.

[expanoncolin]

Gull wing refers to a PDIP-dimension package with legs flattened so that it can be surface mounted.  http://www.analog.com/static/imported-files/packages/PKG_PDF/Pdip_SMD(NS)/NS_8%20(4).pdf

[ve3wwg]

I should have known that, doh!  The description is apt, now that I think of it.

Assuming otherwise the normal DIP dimensions, that might be a preferred format for dead bug style projects actually. 

[ve3wwg]

Understanding takes time, patience and some experimentation. Our experiments confirm that our understanding is correct (or not). Sometimes actual circuits bring in other factors that we've not thought about.

Did you have a specific area that you wanted to ask about? Maybe we can help.