MXR 118 Analog Delay with PT2399

Started by Mark Hammer, September 19, 2007, 02:54:43 PM

Previous topic - Next topic

Mark Hammer

This is a call-out for some assistance in solving a noise issue that prevents a delay pedal from being usable.

Some of you may be aware that NOS MXR boards and other stock sitting in long-term storage were purchased by a number of individuals, and placed for re-sale, some of it here and some on E-bay.  Among the various items were the old 118 model green analog delay pedal.  In its various incarnations, this used a trio of SAD1024 chips and eventually a 4096-stage Reticon R5101.  Given that the basic design had some nice features (a tracking lowpass filter that adjusted treble-cut with delay time), it was tempting to want to figure out some way to make the pedal live again, even if the requisite R5101 chips could not be sourced.  The "solution" was to retrofit a small board with a PT2399 digital delay chip covering approximately the same delay range.  You can see the way the board was implemented in these two pictures, sent to me by Erik in Italy (the person with the misbehaving pedal).

http://img.photobucket.com/albums/v474/mhammer/MXR118.jpg
http://img.photobucket.com/albums/v474/mhammer/MXR118flipside.jpg

The noise being made is best-described as a kind of "thwup-thwup-thwup" helicopter-like noise whose pitch and frequency corresponds to the delay time setting.  Erik sent me soundfiles for diagnostic purposes, and I think that verbal description pretty much captures it. 

You will note that the original 118 ( http://img.photobucket.com/albums/v474/mhammer/MXRAnalogDelayschem1.png ) used a clever scheme whereby the master clock was divided down one way for the BBD chip, and another way for the switched-resistor lowpass filters.  The filter sections are very much like those used in the MXR Envelope Filter, with duty-cycle determining how long a CMOS switch section stays open or closed on average.

While the original used this arrangement, the retrofitted version cannot...sort of.  Instead, it uses a dual-ganged pot.  One half adjusts the master clock on the board, as per usual, and the other half adjusts the delay time of the PT2399, in exactly the same way it is done in the Rebote, FAB Echo, and similar pedals.  I wouldn't call it "authentic", but the goal is to have more treble cut at long delays and less cut at short ones.  The sound files I heard demonstrate that it works.

Unfortunately there is also this helicopter noise that prevents it fro being a nice-sounding pedal.  Erik did not mod it himself, but bought it as is, and would like to make it usable.  Does anyone here have any experience with these refurbished/modded 118s?  We have considered some of the usual culprits like wire leads to close to the on-board transformer, but to no avail.  This leads me to wonder if the problem is endemic to the retrofit itself.

theehman

I bought one of those boards on eBay already modded but there was no small board.  The chip was attached right to the main board.  I have the mod documentation that came with it but I've never hooked the board up to see if it works.
Ron Neely II
Electro-Harmonix info: http://electroharmonix.vintageusaguitars.com
Home of RonSound effects: http://www.ronsound.com
fx schematics and repairs

TELEFUNKON

heterodyning?
(the original version had 2 locked frequencies from 1 clock, obviously? - the new version got 2 clocks, "synchronized" only by the ganging of a 20% tolerated stereopot?).

Mark Hammer

Quote from: theehman on September 19, 2007, 04:37:10 PM
I bought one of those boards on eBay already modded but there was no small board.  The chip was attached right to the main board.  I have the mod documentation that came with it but I've never hooked the board up to see if it works.
Weird.  There is no way that:
a) it could be an original board and still fit a PT2399
b) a PT2399 could be incorporated into an original board without an adaptor board.

Of course, since Erik bought this from someone who had gotten it from somewhere else, we have no idea if the daughter board is what the  retlrofitted pedals look like, or whether someone had taken it upon themselves to retrofit a board of their own device and wiring.  certainly from the description and pictures, the add-on is not what you'd call "professional".  I just don't know how planned-out or improvised it really is.

Ry

I've got two of these boards (currently chained together for some wonderful feedback/pitch shifting fun).  Neither of mine have this helicopter sound.  If you send me some pictures, I can match it up with the workmanship on mine.  The mods to add these daughter boards not 'professional' (no PCB, just perf board and wires), nor are they sloppy, they look much like what a well-experienced builder would do.

Let me know how I can help out, these things sound really nice.

Ry

slacker

Mark, do you have a schematic for the PT2399 daughter board or a better picture of the daughter board. Just looking at the photos there don't appear to be many components on the board and I'm wondering if there isn't enough filtering going on as that can lead to weird noises.

Mark Hammer

Unfortunately, no schematic.

What I can say is that the thwup-thwup-thwup noise not only fluctuates in pitch with ddelay time (higher for shorter delays), but also appears to be synced to delay time (i.e., higher-pitched faster noises at short delays).

The thing with the 118 is that it had a lot of lowpass filtering.  Not necessarily more stages than other comparable contemporary pedals, but the filter corner-frequency was set to  adjust to delay time instead of being fixed and striking a compromise as most other pedals of that era did (and many still do).  So audio filtering should not be the source.

On the other hand, the audio filtering itself is a clocked/switched function, which may be the source of the noise.  Then there is the small matter of what filtering needs to be provided for the PT2399 specifically versus what serves as adequate filtering if there were an R5101 sitting in the slot.  If you look at the datasheet for the 2399, the daughter board looks a little on the "lean" side in terms of components, but I can't tell what aspects of the appnote basic application are being well-served, and which are being under-served.

slacker

I've had another look at those daughter board photos and even assuming the original circuit is providing enough filtering of the audio signal there's doesn't appear to be enough stuff on there for the chip to work properly.
I've been messing around a lot with PT2399s recently and for them to work I'd say you need at least, the following.
Big decoupling caps 47uF or so from pins 1 and 2 to ground and 100n caps from pins 7 and 8 to ground. Without these you get the sort of noises you've described, without a decent sized cap off pin 2 the internal bias for the chip fluctuates and you get a lot of noise. The board looks like it has caps in those positions but I can't see the values.
Pins 9-10 and 11-12 need 100n caps between them. Pins 16-15 and 13-14 are inverting opamps so they need input and feedback resistors and caps in the feedback loop. Without caps in the feedback loops they soon go into self oscillation no matter how much external filtering there is.
On the photos there's only 2 resistors on the right hand side of the board and pins 13-14 don't appear to have a cap between them, so I can't see how the chip can be working properly.
Any chance you could get decent front and back shots of the board and maybe some details of where it's connected to the original board?
The other thing I'm wondering about is where the 5volts to power it is coming from, there certainly doesn't appear to be a regulator on the daughter board, and I can't see 5volts on the MXR schematic.

theehman

I promised this a while back but I never got the chance to do it till now.  This is a PDF of all the data I got from The Gear Trader when I purchase an MXR Analog Delay board modded w/ PT2399 from them.  I've not had a chance to hook the board up so I don't know if it has the same issues or not.  The PDF contains:
schematic for SAD5101 model (2 pgs: one original, one showing changes)
Calibration for original units and PT2399 units
test hookup for PT2399
wiring diagrams

Some of it is a bit difficult to read but I cleaned it up as best I could.  Hopefully the info can be used to help come up with a better mod to use the PT2399 in  place of the SAD5101.

http://www.ronsound.com/ebay/MXRPT2399.pdf  File size: 647kB
Ron Neely II
Electro-Harmonix info: http://electroharmonix.vintageusaguitars.com
Home of RonSound effects: http://www.ronsound.com
fx schematics and repairs

Mark Hammer

Here's a better picture of what Erik sees on his pedal.  Looks to be identical in components to what Ron's PDF  (thank you, thank you thank you!) shows.



So any clues as to what the problem described might be, given that circuit?

bioroids

Eramos tan pobres!

Mark Hammer

He sent me a sample, but I'm not sure how to post soundfiles.

Rob Strand

Helecopter noise could be PSU bypassing, long ground wires, or long clock wires.   See if you can separate your analog and digital grounds and supplies.
Send:     . .- .-. - .... / - --- / --. --- .-. -
According to the water analogy of electricity, transistor leakage is caused by holes.

Barcode80

just an FYI, virtually EVERY one of these floating around were modded to the pt2399 by the gear trader. i bought 3 or 4 of them from Mr. Huge's storage locker. I'd be interested if we could make this work, as i have numerous fully populated boards of these. mine looks just like the pic.

apparently the gear trader bought the original mxr liquidation lot and had planned on reviving them all and repackaging to sell, but decided the time investment wasn't worth it and began unloading the newly modded boards. Mr. Huge still has a ton, as i understand.

domenico

Hi Mark,

I was wondering about the possibility of replacing the reticon 5101 with 4 mn3008 ics in a  MXR 118 Analog Delay .

If you look at these schematic of mxr chorus schematics attached (one with 5101 the other one with a single mn3008) you'll see what I mean :

http://i1206.photobucket.com/albums/bb457/mastrumisciu/mxr_stereo_chorusmn3008.jpg?t=1292252585
http://s1206.photobucket.com/albums/bb457/mastrumisciu/?action=view&current=MXRStereoChorusmodel134reticon5101.jpg

What do you think about it ?

Best regards,

Domenico





Quote from: Mark Hammer on September 19, 2007, 02:54:43 PM
This is a call-out for some assistance in solving a noise issue that prevents a delay pedal from being usable.

Some of you may be aware that NOS MXR boards and other stock sitting in long-term storage were purchased by a number of individuals, and placed for re-sale, some of it here and some on E-bay.  Among the various items were the old 118 model green analog delay pedal.  In its various incarnations, this used a trio of SAD1024 chips and eventually a 4096-stage Reticon R5101.  Given that the basic design had some nice features (a tracking lowpass filter that adjusted treble-cut with delay time), it was tempting to want to figure out some way to make the pedal live again, even if the requisite R5101 chips could not be sourced.  The "solution" was to retrofit a small board with a PT2399 digital delay chip covering approximately the same delay range.  You can see the way the board was implemented in these two pictures, sent to me by Erik in Italy (the person with the misbehaving pedal).

http://img.photobucket.com/albums/v474/mhammer/MXR118.jpg
http://img.photobucket.com/albums/v474/mhammer/MXR118flipside.jpg

The noise being made is best-described as a kind of "thwup-thwup-thwup" helicopter-like noise whose pitch and frequency corresponds to the delay time setting.  Erik sent me soundfiles for diagnostic purposes, and I think that verbal description pretty much captures it. 

You will note that the original 118 ( http://img.photobucket.com/albums/v474/mhammer/MXRAnalogDelayschem1.png ) used a clever scheme whereby the master clock was divided down one way for the BBD chip, and another way for the switched-resistor lowpass filters.  The filter sections are very much like those used in the MXR Envelope Filter, with duty-cycle determining how long a CMOS switch section stays open or closed on average.

While the original used this arrangement, the retrofitted version cannot...sort of.  Instead, it uses a dual-ganged pot.  One half adjusts the master clock on the board, as per usual, and the other half adjusts the delay time of the PT2399, in exactly the same way it is done in the Rebote, FAB Echo, and similar pedals.  I wouldn't call it "authentic", but the goal is to have more treble cut at long delays and less cut at short ones.  The sound files I heard demonstrate that it works.

Unfortunately there is also this helicopter noise that prevents it fro being a nice-sounding pedal.  Erik did not mod it himself, but bought it as is, and would like to make it usable.  Does anyone here have any experience with these refurbished/modded 118s?  We have considered some of the usual culprits like wire leads to close to the on-board transformer, but to no avail.  This leads me to wonder if the problem is endemic to the retrofit itself.

Mark Hammer

Unfortunately, with photobucket blocked at work, I can't think anything about it.

In some respects, you can replace any BBD circuit with any other BBD circuit, provided that:

  • all the clocking is adapted
  • all the polarity/orientation is adapted
  • the filtering is suited to the clocking that was implemented

So, yeah, you could probably do it.  Would it be worth the trouble?  I don't know.

domenico

Quote from: Mark Hammer on December 13, 2010, 10:14:53 AM
Unfortunately, with photobucket blocked at work, I can't think anything about it.

In some respects, you can replace any BBD circuit with any other BBD circuit, provided that:

  • all the clocking is adapted
  • all the polarity/orientation is adapted
  • the filtering is suited to the clocking that was implemented

So, yeah, you could probably do it.  Would it be worth the trouble?  I don't know.

Thanks,
I've also sent you a private email with schematics attached
Regards,
Domenico

DJGlukBH

I'm sorry about the necroposting... But anyway - got the M118 unit with dead R5101 (heats up when the device is turned on, eh).
I'm going to perform the PT2399 conversion. But I'm a bit stubled upon the 78L05 usage for its powering. Wny not to run in on 15v? It will sound damn better! Anyway, I'll keep the thread posted with my findings...
Thanks, Vasiliy

Mark Hammer

If a PT2399 sounds "better" using 15V for you, you'd be among very exalted and rare company.  The PT2399 normally runs off +5V.  I've never seen a schematic using it that did NOT run it from 5V, and the data sheet indicates 5V.

Now, the rest of the M-118 may well like 15V, but the PT2399 sure as hell doesn't.  Use a regulator to drop that 15v down to 5 for the chip.

mistahead

Did the bloke who did this mod remember there is the D-GND and A-GND? Someone also suggested similar back up a bit...

I recall it took forever for some of the brains around here to realise that was impacting on PT2399 circuits, in no small part because of the internal resistor that was omitted in later batches of the chip.

I'm still a newbie.