THE ENGINEER'S THUMB... At last, a better compressor!

[garcho]

^ did you do a continuity check with your DMM?

[InLimbo]

With work and school and all, I haven't been able to touch it, but will definitely have a chance tonight. The idea of the input being sent straight to ground hadn't occured to me until the following morning, and I haven't had a chance to do any continuity tests on the input between the switch and on board. I know the switch has continuity where it should as I've tested that already, but that's about as far as I've tested the input.

I initially thought it was something to do with the components, hence the reason I posted the voltages. Since I've found that the voltages check out, that points even more to the input signal getting lost. I'm really hoping that I'll have it working tonight.

Again, thanks for the assistance. You guys are great, and I'm happy to have joined the forum.

[InLimbo]

CRISIS AVERTED, GUYS.

The absolute smallest solder bridge between the input row and ground. Works perfectly now.

Thanks for all of your help! Now to get the enclosure ready!

[MrStab]

i'm building another one of these to give to some undetermined person (ie. i'm bored and they're fun to make). probably a dumb question, but i ordered a bunch of these LEDs hoping for the most "standard" rating of red LEDs, and i was wondering if anyone could let me know if they'd be suitable for achieving the 3.2Vp-p input limit to prevent clipping: http://www.rapidonline.com/electronic-components/kingbright-l-7113id-5mm-2v-red-diffused-led-45mcd-55-0155. i think i gather that the forward voltage is 2V, reverse is 5V (?), but in all honesty i don't see the number 3.2 anywhere so thought i'd ask! lol
i omitted them from my own build as i knew it'd always be first in my chain and the signal would be nowhere near the headroom limit.

cheers

[psychedelicfish]

If you look at the datasheet for those LEDs, you'll see a graph titled forward voltage vs. forward current. You'll notice that there is a smooth curve starting at about 1.7V. When the input signal reaches 1.7V the LEDs will begin to conduct, and as the voltage increases, the LEDs will conduct more. This results in a soft clipped waveform, and the 3.2V is just an estimation for the peak to peak voltage of this clipped waveform. The reason for these LEDs is that if you feed a large signal into the ET, the LEDs will clip it softly, rather than the OTAs hard clipping it, which means it wont sound absolutely disgusting if you momentarily overdrive it.

So, in short, yes those LEDs should be fine.

[MrStab]

thanks for the explanation, Edward - i think i get now that it's not such a sudden cut-off from the forward voltage, as it were. there were red LEDs with different voltages marked in the title, which made me paranoid in case i'd bought some special niche one lol. either i overthink or i think in the wrong areas.

all i'm lacking are a bunch of 10k resistors, and since i slept through the post this morning, 100 of them will be stuck in a warehouse until monday, when i'm busy with band stuff. this is most uncool. hmm, there are only three 10k's in the circuit, maybe it wouldn't look too atrocious to stick together some 4.7ks or something. closest values in the region i have are 8.2 and 15k.

damn, just realised a few caps i don't have were in that order, too. worst day ever! lol

[garcho]

^ just go to Radio Sha...  oh, yeah, never mind 

[jtn191]

I built mine up and I get a volume boost and more compression as the ratio pot is turned...is that normal?

[midwayfair]

I built mine up and I get a volume boost and more compression as the ratio pot is turned...is that normal?

Yes ... More resistance in the feedback loop.

[MrStab]

Edit: ignore this entire post - upon fixing an error within the ET circuit itself (that i should've noticed) and trying a different batch of 2n7000s, it works. well, at least i managed to track it down by thinking out loud lol. cheers anyway!

my latest build of this seems to work great - but i'm having flashbacks to the weird Millennium Bypass issues i had with the original! around the time i seemed to have bad luck with MOSFETs generally so i phobically switched to the JFET Millennium 1 for a while. haven't tried that with this yet. i'm posting this here because i can't help but think my problem is at least partly to do with resistance to ground on this particular circuit. everyone else's seems fine, so i want to work out what i'm missing.

i get the Millennium working independently, ie. grounding the control line switches it off, but when wired to the DPDT switch, it would only go off if the volume pot was turned CCW slightly. Then, after clicking the DPDT once or twice, it seems the MOSFET dies and it needs replaced. maybe using a NPN as a low-leakage diode would help with the sudden death?

i've tried a coupla 1n4148s to rule out leakage issues. 2 in parallel doesn't seem to help. adding a 2.2M resistor to ground after the volume pot (which might screw up the taper, no?) doesn't work.

sorry for yet another redundant issue on this topic, in my defence i'm just a fan of the Engineer's Thumb! lol. "I built a kick-ass compressor in a day... but i can't get the light to come on" doesn't have such a great ring to it.

[MrStab]

observation: i've built two of these now, and both have shown issues with ticking/LED flashing when bypassed, unless you ground the input. easily done by shunting the signal via. a NPN, but worth noting.

The extra controls are not terribly useful ... there will be some "broken" settings on the attack and decay controls

( http://www.diystompboxes.com/smfforum/index.php?topic=101630.msg898103#msg898103 )

i've also noticed the above, and the ticking is affected by the Release pot. i wonder if a combination of using all 5 controls (or just deviating from stock values) and everyone else using 3PDTs which ground the input anyway is why no-one else has noticed the ticking.

[jtn191]

i've also noticed the above, and the ticking is affected by the Release pot. i wonder if a combination of using all 5 controls (or just deviating from stock values) and everyone else using 3PDTs which ground the input anyway is why no-one else has noticed the ticking.

In my build, I don't currently have any problems with ticking or "broken settings" on the controls. I forwent the millenium bypass circuit and used a 3pdt, grounding input on bypass. I did end up reverse wiring the attack and threshold controls because they operated backwards from how I wanted/expected them to...counter-clockwise should be a short attack time, meaning reduced attack volume and a low threshold should mean more compression for a given ratio...could be my mistake drafting the pcb layout

However, I'm having problems on my second build. This one is a two knob control scheme and I'm experiencing much lower volume output. I'm thinking it may be that I included the 220k resistor near the 1M release resistor...the 220k isn't present on the two knob schem. My 4.5v test points are sitting at 4.8v

[merlinb]

I'm thinking it may be that I included the 220k resistor near the 1M release resistor...the 220k isn't present on the two knob schem.

That's unlikely to be the problem. You want the total release resistance to be somewhere between 220k and 1.5M. That's why it is there in the 5 knob version- it stops the release resistance from being turned all the way to zero ohms.

My 4.5v test points are sitting at 4.8v

Suspicious...

[jtn191]

Found the problem! My 220r resistor was ~2.79k

[jtn191]

Really enjoying your preamp book, Merlin! I got it for Christmas. Reading the explanations and alternate solutions is seriously addicting

I was just checking to see if I had my release control wired properly when I decided to do a little analysis of the range of attack times on the E.T. I found it to be ~4ms-160+ milliseconds. Higher attack times are tricky to measure, so I wouldn't be surprised if it's larger!

[kbibs]

Hi Everyone and Happy New Year!

I built the two knob version of this on a prototype board and it sounds really nice. I like it alot, thanks Merlin! ... Before building it into a pedal (actually the 5knob GT1 version I plan on building) I wanted to play around with decreasing the release time. So I lowered the 1M release time resistor down to 220k and now I hear a ticking noise, almost like static, after the initial attack say during the release and for some time after. It definitely reduced the release time but now with this noise. I then increased it to 511k and the amount of ticking decreased but its still there. With 1M I don't hear it. Like I said I only hear it after the initial attack and usually after play a chord (stronger signal maybe?) rather than single notes. I built it with the led protection diodes and without the Millenium Bypass. Other than this the compressor is functioning as expected. This sounds similar to what some of you have experienced so I was wondering if anyone has found what is causing this? or what they changed to eliminate it? ... I plan on triple checking my component values and maybe trying some different diodes but other than that I'm not sure what to try.

Thanks,
Ken

[kbibs]

Found it!!! ... the second 1uf cap after the rectifier was not grounded. It's very quiet now!

Thanks,
Ken

[NCSUZoSo]

Out of curiosity, what is the build cost for this pedal (not including enclosure, but everything else)?

Also I have seen a 5 knob version of this, how much does that affect cost?

[waltk]

Out of curiosity, what is the build cost for this pedal (not including enclosure, but everything else)?

Like all DIY pedals, it costs every bit of your spare time, many dollars to buy all the tools and supplies you need (and some you don't need), and possibly your sanity...

Real answer: depends on what you already have, and how you build it.

[merlinb]

Out of curiosity, what is the build cost for this pedal (not including enclosure, but everything else)?

Looking at the BOM for the basic version:
17 resistors, $0.17
8 capacitors, $0.8
3 ICs, $2
2 transistors, $0.2
8 Diodes, $0.08
Sockets and footswitch, $10
Total: $13.25

Call it $20 since I'm working in dollars, which are foreign to me!  

I would guess the 5-knob version would add about $15 on top of that, at most.