Review of delay pedals in GP

[Mark Hammer]

The current newstand issue of Guitar Player has a delay-pedal roundup/shootout/review, covering the:
Diamond Memory Lane,
Dytone Di'Ana,
Homebrew Mimic Mock 1,
Line 6 Echo Park,
Maxon AD-999,
Moogerfooger 104Z,
Onerr DGD-2,
Rocktron Short Timer,
Smart People Factory I-5

As is their style nowadays, they flip the hood up to see the insides of he pedals.  Boy oh boy that Maxon has a bunch of chips and trimpots.  The article describes them as "proprietary", but I gather they are likely Beilings with a house number.  (My guess is that a phone call somewhere had a question like "What ARE those chips?" and an answer like "Oh those? We had them made for us.")  The Moogerfooger, Diamond, and Line 6 clearly have the most extensive sets of features, though not identical ones.  The nice thing to see is that there are a lot of interesting fresh ideas about using delays out there.

Finally, the same issue has a review of the Z-Vex Seek Trem and Jonny Octave.  Obviously a positive review, and no less than we would expect from ZV.  I have to say that although I have always liked the colours and general visual texture of Jason Myrold's artwork on ZV pedals, I increasingly find the icons used for legending cute and humorous.  Th "bypass" icon beside the footswitch for the Jonny Octave has a dot and an arrowed line that swerves around the dot.  Bypass!  The bright switch on the Nano-amp has sunshine to indicate bright.  I wish traffic signs were as universal in their comprehensibility as Jason's icons.

[gaussmarkov]

Boy oh boy that Maxon has a bunch of chips and trimpots.  The article describes them as "proprietary", but I gather they are likely Beilings with a house number.  (My guess is that a phone call somewhere had a question like "What ARE those chips?" and an answer like "Oh those? We had them made for us.")

yes, i read that and wondered.  would someone please explain what "Beilings with a house number" means?  google was no help. 

[Mark Hammer]

Matsushita/Panasonic, the original manufacturer of the "MN3xxx" series of BBDs no longer makes them.  Beiling does, however, and Small Bear sells them.  These are the chips with a BL prefix.  To the best of my knowledge, these are not a "different" chip in the same way that a dual op-amp is not a dual op-amp is not a dual op-amp.  It's the same chip, just made by somebody different with a different prefix on the part number.  Now, if we gave the exact same recipe for brownies to 20 grandmothers with 20 different stoves, even if we supplied them all with the same raw ingredients, chances are the brownies would come out just a wee bit different.  Consequently, I'm not going to proclaim they are identical to Matsushita chips, but they sure can't be that far off.

A "house number" is simply a number given to a part that a corporate client has purchase in bulk from a supplier.  For example, the chip labelled EH1048 on a great many Electro-Harmonix circuits is simply a CA3094 with their "house number" printed on it.  I suppose it is *possible* that a part with a house number could be built to different specs, or end up with slightly different specs, like the brownies, but quite often they are just the same damn part, with a special number for inventory control purposes. 

In this case, there is absolutely NO advantage to paying a company to make you a special BBD that you have to use 8 of to achieve any useful delay length.  Why not just pay them to make you some 4096-stage devices or even a, gasp, 8192-stage chip (there has never been one to my knowledge)?  Crazy to go to any great lengths to have a chip made which requires that degree of fine tuning.  Three SAD-1024's in the old MXR analog Delays was bad enough.  Eight BBDs is just plain nuts.

Nope, my money is on Maxon paying Beiling to crank them out an assured supply of 1024 and/or 2048-stage chips with a house number that distinguishes them from those made for general retail distribution.

Proprietary, on the other hand, would be something like the Rohm BA662, or IR3109, which Roland had made for them specially and exclusively.  Although I can't say what happens behind closed doors in the basements of the Akihibara, mere mortals cannot buy these chips from a distributor.

[gaussmarkov]

thanks, mark.  very interesting.  and it makes sense, too! 

[BDuguay]

I was fortunate to have the oppurtunity recently to audtion the Memory Lane, J Drive, Compressor and Fireburst from Diamond Pedals and let me tell you in emoticons my reaction and experience
B.

[Mark Hammer]

One of the members from the Halifax area had some nice things to say generally about Diamond (they're a N.S. company).  Nice to see a Maritime product hit the glossy pages for a change.

[Peter Snow]

Yea, but they are pricey.  The Memory Lane is going for CDN$450.00 at Song Bird in Toronto.

Peter

[BDuguay]

I agree Peter. I was a little surprised at the price of the Memory Lane too. It will definitely appeal to a select group. However, considering it is a 'Boutique' type pedal, I think the price is in line. Having seen and sampled this pedal I can tell you the overall build quality is superior and it delivers what it promisses. I use delay in my band and personally, this pedal is not for me. But if I had the money I'd buy one in a 100ms heartbeat.
I'll admit that it has something to do with national pride too.
B.

[Melanhead]

One of the members from the Halifax area had some nice things to say generally about Diamond (they're a N.S. company).  Nice to see a Maritime product hit the glossy pages for a change.

I can't say enough about 'em ... I have the J-Drive, Memory lane and the Opto based compressor... Customer service is great too ... My Memory lane pedal died a few weeks after I got it ( apperently there were a few bum chips that after about a week or 2 would just stop working ) anyways, I brought it back to the music store for repair and got a call 3 hours later. Michael Knappe, founder of Diamond, had picked it up and fixed it. I picked it up at his house and he gave me 2 extra MN3005 chips just in case. Now that's customer service! ...

[Mark Hammer]

Small IS beautiful, isn't it?

Your comments made me wonder, though.  Are there runs of MN3005's that are subject to failure, or were these just a special "thank you" because he knows your pedal-buildign habits?  We should squash any rumours right now if at all possible.

[Melanhead]

Small IS beautiful, isn't it?

Your comments made me wonder, though.  Are there runs of MN3005's that are subject to failure, or were these just a special "thank you" because he knows your pedal-buildign habits?  We should squash any rumours right now if at all possible.

Not sure on that one. Michael mentioned that he had 3 out of 350 Memory Lanes ( I'm assuming that's how many have been built to date ... ) stop working in the first 2 weeks and it was always the chip ... I think the extra chips were for my troubles and 20 years down the road, they may be hard to find ....

[BDuguay]

Melanhead,
Describe for me please, your take on the J-drive. Specifically the overdrive channel.
Thanks,
B.

[Melanhead]

Melanhead,
Describe for me please, your take on the J-drive. Specifically the overdrive channel.
Thanks,
B.

Hmmmm, sounds similar to my modded TS808 clone ... It's tubescreamer like but fatter with 2 types of clipping ( compression settings on a switch )  .... Not sure what the op-amp is as it's sanded down for anonimity ... I may someday reverse engineer it out of curiosity but probably won't share out of respect for the manufacturer

The clean boost is great as well, although it's in front of the overdrive section so it adds more gain ( distortion .... )  rather than volume. I'd prefer it the other way around, and may mod it to do so, as it works great as an indpependent solo boost for the pedals I have in front of it.

Overall it's a great sounding pedal ... It's on my pedal board and I've been using it since January.

[Paul Marossy]

Th "bypass" icon beside the footswitch for the Jonny Octave has a dot and an arrowed line that swerves around the dot.  Bypass!

Yeah, I also thought that was very clever. 

[cd]

Nope, my money is on Maxon paying Beiling to crank them out an assured supply of 1024 and/or 2048-stage chips with a house number that distinguishes them from those made for general retail distribution.

You know I wonder what's inside the AD-999 - on the Maxon website they call the eight chips "MC4107D" which leads one to believe they're 1024 stage chips (the number 7 alludes to MN3207 or MN3007).  The now discontinued AD-900 had either two MN3005s (8192 total stages) or four MN3008s (8192 total stages) with an advertised 600ms delay time.  The newer AD-999 has eight MC4107D chips with an advertised 900ms delay time.  If 600ms requires 8192 stages (at a clock speed of oh, say 8kHz or so) then to get 900ms out of 8192 stages (if a MC4107D has 1024 stages) would mean clocking down to well below 8kHz (each stage requiring 112ms, and the 3x07s are rated for 51ms at 10kHz).  A low clock frequency means heavy filtering, which would support the lack of highs reported in the GP article.  Then again if those eight chips are 2048 stages each, then the clock could be right around 10kHz, which would leave plenty of room for highs.  HMMM.

Then again, check out the gut shot of the AD-999:

http://www.maxon.co.jp/MaxonAnalogDelayAD999/ad999_inside542.html

Three clock driver chips marked MC4108D - if they're MN3102 equivalents, each of those can drive up to 4096 stages, so if there were only 8192 total stages, then only two would be required.  But then if there were 16k stages, four would be required.  So why are there three?  Argh...

[Mark Hammer]

First off, let's give credit where it is due. When was the last time you saw a pedal with twenty four trimpots?  Rather than working stuff out and throwing a 5% fixed resistor solution at it, Maxon has seen fit to adjust everything to some sort of perfection.  That's certainly a big chunk of their retail price.  Not just the trimpots, but the obligatory chassis size, and the tweaking...oh the tweaking.  I have visions of some sweatshop where dozens of Koreans or Taiwanese are kept shackled to their benches with a scope, some breadcrusts, and a bowl or water, while an "administrator" with a large drum (like on slave galleys" pounds a beat as the boss yells "Tweak! Tweak!  Tweak! If you ever want to see your children again, TWEAK!!"

But you're right, the math is a little funny.  I was looking through the specs of the Matshita BBDs last night, in an effort to try and understand the clock pin input capacitance, and the deal is that this capacitance is essentially the sum of all the little caps inside that are tied to the clock pin.  So, more stages=more capacitance within the chip, and more chips = more capacitance for the clock driver to see.  The clock capacitance is important to sound quality, if I've understood it correctly, because as the clock pulse is "rounded" by the capacitance (think of it like a treble cut filter imposed on a square-wave input), the clock starts to behave as if it were less than 50% duty cycle.  When the clock pulse is nice and square, then the two complementary outputs trade off seamlessly and you experience the BBD output like a movie in the theatre.  When the clock pulse is shaved a bit, then it behaves as if it has less than a 50% duty cycle and the ever-so-brief gaps between the alternating complementary signal samples arriving at the output start to make the "movie" into aTV show, with ads and promos woven throughout.

The use of multiple clock drivers is a way of keeping the number of teeny capacitors seen by the driver to a reasonable minimum, and thus delivering as close to a flawless 50% duty cycle as possible.  What I can't quite figure out is why there are 3 %^&* chips for 8 BBDs.  Given the spacing, my hunch is that the rightmost two BBDs in the picture linked to (you'll need to enlarge it a bit) are clocked by a discrete circuit just below them on the board, and that discrete circuit synchronizes the other 3 BBDs so everyone is in step.  Thus no clock sees more than 2048 stages, when BBD clock input pins are paralleled.

But why not just use a single clock driver and a buffer?  I gather this MAY (I'm just over the perimeter of my expertise here) have something to do with capacitance arising from long traces, as well as any clock noise accruing from having clock traces running all over the place.  Normally, BBDs are snuggled right against their clock drivers to keep those lines short, and this pedal is no different.  Of course it bears remembering that this all seems to occur because of the decision to use 8 smaller capacity chips, rather than a few higher capacity ones.  It is quite possible that the low capacity ones have the potential for lower noise or distortion.

Those green things that all say "CP" beside them are test points, I gather, for verifying the needed voltages.  This is clearly a meticulously set-up pedal.  The use of trimmers on the BBD output (and I gather this is the row closest to the BBD's where the trimmers are all set to about the midpoint) instead of approximately "matched" fixed resistors as many traditional analog delays use (e.g., Boss DM-2) comes closer to having the ideal seamless signal emerging from the BBD.

What you DON'T see on this board is evidence of a lot of filtering.  I see evidence in the upper right hand corner of what is likely some fairly standard 2 or 3-pole lowpass filtering around transistors.  The number of resistors in evidence,and the physical size of the caps syggests that's where the lowpass filtering lives.  But there is NO evidence of companding, unless there is another board hidden underneath.  This is probably why the need to fine tune both the bias and balance of each BBD: to get absolutely maximum signal level into those babies, so that max S/N ratio can be gotten without companding.

And just so you know, I did a word count on this posting, and it seems that in this case, a picture was worth 783 words.

[cd]

Hmm all those trimmers threw me a bit too.  You'd think it would be cheaper/easier to match all the BBDs beforehand and then throw a bunch of fixed resistors at them.

FYI I think the AD-999 does use a compander, there's a long chip underneath that turning point of 8 wires (1 orange and 7 white).  I can barely make out the Philips logo, and given the number of pins it's likely a compander.

[Mark Hammer]

I was thinking that too, but you obviously need an input stage, mixer for handling regen, and an output mixer stage, so without any other contenders on the board, the odds are its a quad op-amp.  Again, all bets are off if there is another daughter board underneath the one we can see.

[Tim L]

I was thinking that too, but you obviously need an input stage, mixer for handling regen, and an output mixer stage, so without any other contenders on the board, the odds are its a quad op-amp.  Again, all bets are off if there is another daughter board underneath the one we can see.

Hi Mark,

Think there's any chance the 3102s are in parallel to increase the driving capability? I thought it was strange that 3 clock chips are used to drive 8 BBDs as well, thought about opamps being put in parallel (or stacked) and wondered if maybe that's an option for clock chips. I'm hoping to get my hands on and ad-999 to find out.

Hope all is well!
Tim

[A.S.P.]

at 20kHz the 3102 has no problems with "rounded corners", but still has to pump an amount of current to load/discharge 1000s of caps...
[just a note]