SAD1024 alternatives?

[finkfloyd]

Hi is there any alternatives to these hard-to-find IC's? any substitutes?

I would like to build an electric mistress or univibe, but these chips are like trying to find rocking horse shit, or they cost more than the actual pedal is worth!
I found a site in the UK selling these for £50 each, I believe the pedals £90 new

Thanks in advance.

[nelson]

unfortunately not, there are no pin for pin SAD1024A equivelants. However Smallbearelec.com has them at a reasonable $14.95. I havent found them cheaper ANYWHERE else, believe me I have been looking.

[StephenGiles]

Think of it like this, in the UK 20 cigarettes can cost £5.50 for 20, so 2 packs cost £11 which is around $17. One SAD 1024 from Steve Daniels costs $15........need I say more!!!!
Stephen

[finkfloyd]

Thanks for the response guys, I will get some from him after christmas!

PS: Where are electro harmonix getting these from as sovtek dont seem to stock them, and i thought they would be the obvious suppliers?

They must be like area-51, making their own in secret...

[MR COFFEE]

IMHO,

There is *no* substitute for the SAD1024A.

It is an n-channel device; the p-channel devices (Panasonic MN seriesd and various clones and work alikes) do not behave as well as n-channel devices as you shift clock frequencies up really high.

The bias needs shift significantly at higher clock frequencies in the p-channel devices - and p-channel devices just don't behave as well when the clock gets up above 125 khz.

$14.95 is cheap!

[Steben]

The bias needs shift significantly at higher clock frequencies in the p-channel devices - and p-channel devices just don't behave as well when the clock gets up above 125 khz.

Euhm...
What the h**k am I doing then to my BF-2
The original clocking already goes from 20kHz to 500kHz, and I am even going over that.
Check Mark Hammers notes on this topic.

[MR COFFEE]

Hi Steuben,

It'll work, as you know. But look at it on a scope with a triwave input set at maximum undistorted level at 20 khz and then watch it as it sweeps up to 500 khz. and you'll see what I mean. Course distortion doesn't have to sound bad when guitar is involved

[DiyFreaque]

MN3XXX data sheet specs are based on their respective Panasonic clock driver IC's, the outputs of which start to distort past a certain frequency range.  Using a higher freq clock and buffering the clock signals rather than just using an MN310X is the key to successfully driving the MN3XXX's beyond the datasheet specs.

The later versions of the A/DA Flanger clocked the MN3010 in excess of 650 kHz.  Legend has it the Modcan Flanger, designed by Mike Irwin, goes way beyond that.   I know the A/DA used a clock built from CMOS IC's.  I think I can accurately guess that the Modcan flanger does the same.

Haven't tried it with a 20 kHz input signal, though. 

[MR COFFEE]

Hi guys,

Good points.

I'm not trashing your P-channel devices. I use them in my projects, too. 

I'm just pointing out that the SAD-1024A is unique, and that N-channel devices have positive characteristics that P-channel can only approach - it's basic process physics. Especially clocked fast.

MN3XXX data sheet specs are based on their respective Panasonic clock driver IC's, the outputs of which start to distort past a certain frequency range. 

I didn't know that the MN310x driver chips couldn't put out a decent risetime when run over 100khz; I never tested them like that. I'd be interested in where you got the information that the specs of the drive chip determined the spec of the delay lines - that seems like a peculiar way to spec things to me. Did you get that from somebody credible at Panasonic ? And uh, marketing-types *don't* count in my book

Clock drivers have to drive relatively high capacitance clock lines on BBDs. The clock lines are only 110 pf on a SAD-1024A; P-channel devices of equal length are 3 times higher - the MN3004, a 512-bucket, is 350 pf, the MN3007 clock lines are something like 700 pf IIRC. That's comparable to the gate capacitance of a power MOSFET like the IRF-series used in line power supplies.

Using a higher freq clock and buffering the clock signals rather than just using an MN310X is the key to successfully driving the MN3XXX's beyond the datasheet specs.

Makes sense to me it'd take a good driver. That's a lot of charge to move around quickly at really high clock speeds.

But here's why I say there is no real substitute for the SAD-1024:

SAD-1024A's are *rated* for 1.5 mhz maximum clocks, and in alternate-sampling mode they easily sample over 3 mhz *within* specs. MN series are *rated* 100 khz; that's more than an order of magnitude lower.

Yeah, I know they will work much faster. The clock feedthough is , well, look at it on a scope sometime when you've got a flanger open on the bench. I suspect the 100 khz clock rating has more to do with that, but enlighten me if someone around here used to work for Panasonic and has the inside scoop. I didn't measure S\N at 500 Khz. Anyone bothered? It's 70 db *unweighted* on the Reticon.

Dispersion losses at those speeds ? Slight with the n-channel SAD-1024A. The gain just doesn't drop, even at really high clock speeds.

I agree the Panasonic chips work fine, and they can sound d*mn good, too. No arguement. I use 'em. Does it shorten their life to run them more than an order of magnitude faster than the manufacturer's rated maximum clock speed? Who knows? The heat dissipation should be mostly in the clock driver circuitry, but...

IMHO, if you want really short delay time for a flange or something, why not use less stages and run 'em slower, and get less noise (which goes up with the number of stages), or, if you need the wide range of delay of a really fast 1024-stage device for some reason, just get a SAD-1024A and run it in alternate sample mode *within* spec. 

P-channel are fine devices, but they *ain't* n-channel. I still would say the SAD-1024A is in a class by itself.

[A.S.P.]

3207/-8 is "N-channel"...

[Mark Hammer]

Clock drivers have to drive relatively high capacitance clock lines on BBDs. The clock lines are only 110 pf on a SAD-1024A; P-channel devices of equal length are 3 times higher - the MN3004, a 512-bucket, is 350 pf, the MN3007 clock lines are something like 700 pf IIRC. ......

..... I still would say the SAD-1024A is in a class by itself.

That's why the 3101 and 3102 clock drivers are spec'd for # stages.  In the Matsushita MN30xx, 32xx and 33xx series, number of stages is directly proportional to capacitance; a 2048 stage device has exactly the same input capacitance as a pair of 1024-stage devices in parallel, and half the input capacitance of a 4096 stage device.  The MN3101 is spec'd to directly drive up to 8192 stages (eight MN3007 or two MN3005) within a certain frequency range.

It's not so much that the MN3101/3102 can't put out decent rise times.  Rather, it is that they can't, of themselves, assure that the clock pulse sent is the clock pulse received by the BBD, in the absence of buffering that ups the current enough to overcome the effect of that input capacitance.

I have been a big fan of the Reticon chip over the years, frequently noting that pedals I liked more had used them.  On the other hand, I think it is relatively safe to say that the major difference between the Reticon and Matsushita chips is not so much in what they can accomplish, but in how much help they need to accomplish it.  The SAD-1024 needs much less help in most normal instances, and in that respect it IS in a sort of different class.  A bit like the difference between two vehicles that both have the samerange of speeds attainable, but one accelerates much faster than the other; no perceptible difference at 140mph, but a very perceptible difference getting there.

[R.G.]

The SAD1024 has a higher top speed. One of the things it was used for was correcting video time delays. The Panasonic stuff simply can't do that. It was/is a unique chip.

[stm]

The SAD1024 has a higher top speed. One of the things it was used for was correcting video time delays. The Panasonic stuff simply can't do that. It was/is a unique chip.

Agreed, however the MN3007 has a much better S/N ratio (80 dB v/s 70 dB), so basically the SAD1024 is inherently noisier and cryes for companding 

IMHO each has its field of usage: I reserve the MN3007 for choruses due to its lower noise and prefer the SAD1024 for flangers due to its increased high-end and max to min clock ratio.

[DiyFreaque]

To paraphrase Elaine Benes of 'Seinfeld' - the circuit has to be 'spongeworthy' in order to commit a SAD1024 to it.

I've always wanted to try the TDA1022 - it seems to fall in the middle between stated Reticon and Panasonic clock speeds with slightly better S/N than SAD, worse than MN30XX.  Looks like it can handle a much larger input signal than either.  It's a 512 stage device, which is something I'd like to have around (why doesn't Belling pump out a 512 stage device?).  Sure is rarer to see it used in effects as opposed to Reticon/Panasonic, though there are certainly effects out there that use them.  I recall Steve Daniels mulling over stocking them at Small Bear.

[StephenGiles]

All the TDA 1022 BBDs I've used have been much noisier than its SAD 1024 brother or sister (which ever takes your fancy - good old Elton!!!) I have a Standard Electric Mistress with "bounce" using a TDA 1022 which is very noisy but interesting nevertheless - one day I'll draw out the circuit. The Carlsbrough Mantis Echo used maybe 4 of them with companding and some of the "simple" units published years ago in Practical Electonics and like magazines - ah yes the first one I ever saw was in Elektor in the late 1970s. But I don't recall any actual pedals which used a TDA 1022.
Stephen

[puretube]

there once was a L*slie-simulator by Dynac*rd, that had 4 x TDA1022 in them...
(not a pedal, though)

[DiyFreaque]

Probably ETI-like projects I recall.  Seems to me the Infinite Flanger project used a couple of them?

Cheerio,
Scott

[MR COFFEE]

Hi stm,

Agreed, however the MN3007 has a much better S/N ratio (80 dB v/s 70 dB), so basically the SAD1024 is inherently noisier and cryes for companding 

Is this an I-measured-it-in-this-circuit S\N ratio?

Comparing the data sheet specs on the two chips is "apples-to-oranges". Reticon gives the straight unweighted S\N ratio: Panasonic runs it through an a-weighted filter which makes it *look* like it is lower in noise.

I've never measured them myself, but I do get the impression that the post-BBD amplifier scheme can make quite a difference in terms of the audio noise as well.

Frankly, I think ALL the BBD devices "cry for companding", but I've kinda got a thing about noise.

[MR COFFEE]

Hi A.S.P.,

Thanks for pointing out the MN

3207/-8 is "N-channel"...

I got in the habit of referring to the Reticon chips as n-channel back in the good-ol-days, and there *are* other n-channel BBD parts out now like the Panasonic MN3207 and MN3208. 

They are built on a different n-channel process than the Reticon devices were, however, and were not built for high clock speeds.

The Panasonic MN3207 and MN3208 are built on a n-channel process that works better than the p-channel devices in low voltage applications, targeting 9 volt battery-operated apps. They won't run at supply voltages over 10 volts, though, and are rated for 200khz maximum clock speed. 

When I want a really shimmering flange, or want to be able to turn the feedback up *way* high for that really intense metallic sound, I've just got to use for a SAD-1024A. They just sound better to my ears for that sound. Panasonic chips get close, but they just don't get *that* sound.

[stm]

Is this an I-measured-it-in-this-circuit S\N ratio?

Comparing the data sheet specs on the two chips is "apples-to-oranges". Reticon gives the straight unweighted S\N ratio: Panasonic runs it through an a-weighted filter which makes it *look* like it is lower in noise.

I've never measured them myself, but I do get the impression that the post-BBD amplifier scheme can make quite a difference in terms of the audio noise as well.

Yes, you may be right regarding the noise specs. Checked again the datasheets and they are not given in the same terms. However the question as to which (SAD1024 or MN3007) has better PRACTICAL noise specs is still open.  I feel that the MN3007 might still be somewhat better than the SAD1024, but probably not by as much as 10dB's as I stated before.

I agree plenty with the fact that the SAD1024's sound better for flanger. This "something" they have over the MN family might have to do with the better bandwidth which in turn produces deeper notches.

Post filtering is important, especially if you are using pre and de-emphasis to further reduce the high frequency noise.