finally finished my PT80

[Brian Marshall]

this thing is pretty cool.  not exactly what i was expecting.

I was really worried on this one that I would have to trouble shoot it, but luckilly the only trouble i ran in to was a wire i forgot to solder, which took no time at all to find.

i made a minor mod to it, by using a 100k pot for the delay time.

this thing seems to be pretty close to 1000ms delay time with out any nastyness.

self oscilation is a little more shrieky than i thought it would be, but i dont really have any use for that anyways

I'll post some clips next week.

and thanks again mark for the regulators

Brian

[muizac]

pretty cool. do you like the sound of it?
I havent seen any instructions for a pt-100 - is there a schem available? I'd be interested to see what chip can get to 1000ms.

[Yuan Han]

i think you meant pt-90 ?

its available at generalguitargadgets, www.generalguitargadgets.com.

using the PT2399 chip its really convenient for the delay time as the delay time is set by resistance. my pt-90 is modded with a short/long delay switch thats switched in by a dpst switch that switches in an additional resistance ( i think 47k) in series with the pot. (ala moos' mod on his website)

Han

[Fret Wire]

PT-80?

http://www.generalguitargadgets.com/v2/index.php?option=displaypage&Itemid=80&op=page&SubMenu=

[Yuan Han]

aha... my apologies.....

its not PT-100 nor PT-90 but PT-80 !
named after ibanez's AD-80 i think.... according to Scott's notes.

Han

[Brian Marshall]

im a dumb ass.... oops....  PT80

anyways, yeah.... it sounds good.

I took a really good listen to the longer delay times tonight.... played clean.  yes it is a little distorted at the longer delay times, im sure if you went any longer it would start to sound really bad.

I timed it tonight, it is more like 1100ms..... cool... it sounds pretty.

now i have a DD3, a rehoused DOD fx90 and the pt80 for delays.... plus some other rack gear, but that is more studioish stuff.

Now i just need to finnish my green ringer..... of the 3 projects i started.....  green ringer, pt80, and dod 440, you would think the green ringer would have been done first...... but nope....  Soon....  :roll:

[Maneco]

i love my pt80 too...i added a triangular lfo,from the memory man delay schematic,with the switch to change capacitors,that drives a led,that modulates an ldr in parallel with the time pot...cool chorus and vibrato...i also wired a 250 k pot in series with the ldr as a depth control...some modulaton always bleed,though...but i love it...

[Yuan Han]

heyo maneco,

for it to have a chorus effect, wouldn't the delay time have to be in the range of say about 20ms ? hence the ldr chosen should have a resistance light/dark resistance value in that range ?

and hmm... is 250k too big a resistance value to be in series ? since I think that it would drive the delay time way beyond 20ms, and hence it would not sound chorusy anymore ?

I was thinking perhaps a 5k pot in parallel with the ldr actually.

but i guess if it works....its fine!

Han

[ExpAnonColin]

I've always meant to get around to this... I have a ton of PT2399's, and plan on building a few things of my own.

-Colin

[muizac]

cool.

i thought the pt-100 was an upgrade

I didnt realise that you could get 1000ms from a PT2399 chip - i thought it was more around 500ms. I've been looking for something that could handle 1000ms as i play dub guitar at around 60-70bpm - and need 1:1 beats - so this could be the one for me. (now very excited)

Are there schematics available for the LFO mod? I was going to build a flange pedal guts and stick this between the delay and feedback loop - but maybe there is an easier way.

With the switchable resistance - can you put a resister in the chain to cut the delay time to 1/4 of the original delay time? ie I could set the max delay time to match my beats at 1:1, so for example at 65bpm i'd set the max delay time to 920ms,  would pressing the resitance switch then cut my delay time to 230ms and get me 1/4 beats? (whilst if the max delay time was set to 800ms, swithing this resister would give me 200ms delays).  If this is how it works then i'd be over the moon, its exactly what ive been trying to work out for my delay pedal - if anyone has any pointers on this i'd be very grateful (diagrams are great as im still a reletive newbie)

thanks heaps - muizac

[Maneco]

heyo maneco,

for it to have a chorus effect, wouldn't the delay time have to be in the range of say about 20ms ? hence the ldr chosen should have a resistance light/dark resistance value in that range ?

and hmm... is 250k too big a resistance value to be in series ? since I think that it would drive the delay time way beyond 20ms, and hence it would not sound chorusy anymore ?

I was thinking perhaps a 5k pot in parallel with the ldr actually.

but i guess if it works....its fine!

Han

i explained it wrong... i use an ldr in series with a 250k depth pot...that series goes in parallel with the time pot (50k on mine) ...the modulated delay produces a chorus effect,even if it's on the 300 ms range...

[Brian Marshall]

I've always meant to get around to this... I have a ton of PT2399's, and plan on building a few things of my own.

-Colin

hey colin.... it is a very worth while project.  lots of stuff in there.... I'm going to eventually go over the schematic with a fine tooth comb soon, and see what if i can have some fun with it later.

I have a long way off idea for the pt2399 as well.... not sure when ill get to it.  I wanted to build this basically just to see what it was capable of.

Brian

[ExpAnonColin]

Definitely man... I'm considering doing something just like the PT80 except with an even more analog-like feedback loop, with the correct LPF value and all... and perhaps even an in-loop 3 band EQ.  We'll see, we'll see. It's good to hear it can get up to 1 second, though.

-Colin

[Brian Marshall]

Definitely man... I'm considering doing something just like the PT80 except with an even more analog-like feedback loop, with the correct LPF value and all... and perhaps even an in-loop 3 band EQ.  We'll see, we'll see. It's good to hear it can get up to 1 second, though.

-Colin

you can get 1 sec delay.... there is a tiny tiny bit of distortion you can hear on the longer delay times.  A lower frequency filter may help quite a bit.  Squashing the delay more might help too.  The filter is the one thing that i think could use some tweaking.

I think i may add an effects loop for the delay as well.  I used to have a digitech processor that allowed me to run the delays thru an envelope filter.... it sounded really cool....

[Mark Hammer]

Unless I'm missing something big time, 1000msec from a PT2399 just seems unlikely.  The PT2399 has a fixed amount of RAM onboard and tends to poop out around 350ms.  I suppose you can tinker with the clocking of it, but the bandwidth would drop noticeably and signal quality too.  

In contrast, the PT2395 lets you use either 64k or 256k RAM chips with it for different delay ranges.  This is identical to the now discontinued Holtek HT8955, which delivered 800msec delay with a single 41256 chip.

Doesn't mean that folks can't cascade these for long composite delays, though.

[jsleep]

This is identical to the now discontinued Holtek HT8955, which delivered 800msec delay with a single 41256 chip

Hi Mark,

I think you may have your information wrong.  The PT32** chips are nothing like the HT chips as far as I can tell.  There is not an option to use external RAM, etc.

JD

[puretube]

Mark is absolutely right about the PT2395...

IMHO, it is possible to even get a 16second delay out of it,
with just a little more memory and a few dividers...

I bet, there are people currently busy on it...  :wink:

Drawback: afaik, it`s outa production (currently)...

[Mark Hammer]

The PT2395 uses external memory, while the PT2396 and 2399 use internal RAM.  The 2399, in particular has 44k of RAM, although I'm never certain from the datasheets exactly what this means.  Normally, one discusses RAM under the assumption that word-size is 8-bits, but the sampling is 10-bit here, so that could be 44k of 10-bit addresses, OR it could mean something less.  Either way, with a ceiling on RAM of 44k samples in storage at any given moment, that does not permit a great deal of bandwidth once you move out of the 300-400msec zone.  Like BBD's, though, that doesn't mean you *can'T* extend the delay time, but you need to accept what it will cost you in bandwidth and additional anti-alias filtering.

The PT23xx series are not pin-for-pin replaceable with the HT8955, so JD is quite accurate there, but they are *functionally* quite similar in that both Princeton and Holtek produced internal and external RAM versions of a 10-bit all-in-one digital-delay chip.  Like the HT8955, the PT2395 lets you use a 4164 *or* 41256 RAM chip for either 200 or 800msec.  I made one of the latter and it isn't half bad.  Could benefit from maybe a little of the companding voodoo that the PT-80 and AD-3208 have, but all in all not bad at all.

What Ton has made me curious about (and he has a way of doing that....) is how one marries up extra memory to the HT8955 or PT2395.  The datasheet would suggest that  the chip pretty much loses its ability to keep track of where the signal is once you get over the 9 address lines and row/column lines onboard.  I gather there is some multiplexing involved?

Heck, even doubling the delay length to 1600msec would be quite useful for purposes of looping.  You'd be surprised how much you can stuff in a second and a half.

[puretube]

Hello Mark: no, I`m not the guy who designed the new E-H 16second delay   :cry:

but: about 20 years ago I played with a little thing (it had inferior AD/DA conversion than Holtek/Princeton/Mitsubishi(!) ),
and (roughly stated) parallelled the data into a second RAM, which was addressed by the same address-bus, but got its "chip-enable"
from a flipflop (divide by 2) from the last address-line...

[ExpAnonColin]

Brian-Equally as cool is PITCH shifters and ring mods... man oh man.  Did you see moose's site?  He did it, might give you ideas.  I'm going to see if I can't get it to run off of a 9v all right...

Mark-The way I've always understood it is that halving the clock frequency is as big of a deal, storage/THD/bias-wise, at any delay amount...  So if he were to halve the fck to 1mhz, then he'd be getting 700ms without a huge amount of quality loss.  Correct me if I'm wrong, though.  I had always thought it tangible to get 1000ms out of 2 of these chips underclocked, it surprised me that youc ould get it all right out of one.

-Colin