Opionions on analog switches vs vacrols

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm probably one of thousands who in times gone by has used VACTROLS such as their vtl5c3 (Newark Reference) to re-route low level audio. For me, I've typically used it in situations where I wanted to control several older style guitar effects pedals that typically use a mechanical SPDT flore stomp button. By leaving the pedal "ON" and using a few VACTROLS and a control voltage, i could easily either route the audio chain through the pedal, or bypass it. Often I'd use some OP amps in the chain, but just as often it wouldn't be necessary. The quick yet NOT instantaneous response of the VACTROLS always made for a smooth pop-free transition.

But now I'm considering an updated version of the above pedal control methodology, and I've become intrigued at the apparent high quality of solid state analog switches such as the ADG1612 (Analog dev. reference). Its very tempting to use these because a single unit contains 4, making its cost/gate about 1/4 of what VACTROLS are going for (unless I get a VACTROL equivalent knock off from alibaba.com).

So do any of you have any opinions here? The off isolation of the AD chips are touted at -70dB (not bad), and while they don't have the slow roll off time I've come to like in the VACTROLS, real mechanical pedal buttons never had ANY such delays, and I'm thinking they may be a better fit. I understand that for controlled analog systems, like building a compressor/limiter, the VACTROLS are still great. But for simple switches maybe they are overkill?
 
No specific first-hand experience here. I suspect the solid-state switches will never be as smooth and click-free as optically coupled devices like the Vactrol. (The updated and improved successors are made by Silonex.) Whether anybody cares about, or even notices, the difference is another question.

Vactrols could be effectively used with program signal levels all the way from transducer levels (a few millivolts) to line levels (tens of volts). Solid state switches will probably require more thought to the voltage levels being switched - I think they are all limited by their supply rails.

The Vactrols et al most DEFINITELY offered the possibility for isolating the control signals (with their related switching noise, ground loops, etc) from the program signals.

Dale
 
I tend to agree dale. A vactrol will never have the added issues of having to add capacitors when powered from a single ended supply. I guess what even got me looking for alternatives was the prices. Perhaps I haven't found the magic stash yet, but except for those possible allibaba substitutes from china, I haven't found any company stateside making a comparable part for less than $4. Once upon a time I didn't care because anything I made was generally just for me. These days I hardly make anything without considering the possibility of marketing, which of course makes any possible savings on parts important. Do you know of any place to get the optos cheaper?
 
I've become intrigued at the apparent high quality of solid state analog switches such as the ADG1612 (Analog dev. reference). Its very tempting to use these because a single unit contains 4, making its cost/gate about 1/4 of what VACTROLS are going for (unless I get a VACTROL equivalent knock off from alibaba.com).
Ooh, cool, Vactrols. Linear response optocouplers. I didn't know about these. The datasheet from excelitas technology is so clear. I had previously found some led-CDS cell optocouplers by chinaledproduct.com but their datasheet was so weird I gave up the idea of using them. The product number is LCR0202 and LCR0203. No charts on the copy of the chinaledproduct.com datasheet I found. So that gives you another source of a similar product, besides 'alibaba".
I backed into these by looking up parts used in 1963 Wurlitzer organs, where an incandescent bulb/CDS cell Wurli made device was used for volume control etc. It struck me as just the thing for current controlled gain amplifiers, with a cds cell in the feedback loop. Apparently according to the wikipedia VCA article (voltage controlled amp) the concept of cds cells in the feedback loop has been used before.
I know from my work in geophysics that switched input high gain channels using relays get a huge pop when the input switches. And I don't expect an analog switch to be a lot better, although with a CMOS flip flop as driver (74HC74 or 4013b) the switching transient might be so high in frequency as to be capable of elimination by RC high cut filter.
 
Last edited:
. . . . except for those possible allibaba substitutes from china, I haven't found any company stateside making a comparable part for less than $4. . . . Do you know of any place to get the optos cheaper?
It appears that Silonex was bought by an outfit called "Advanced Photonix Inc" about a year ago http://advancedphotonix.com/optosolutions/ . If you waste enough time rooting around their website, you eventually stumble across a slick marketing page that brags about how great their analog optocouplers are, but doesn't give a single performance measurement, or model number, or link to any technical info . . . . so I don't know if they intend to actually sell the things, or just use the web page as vaporware to keep the stock price up and fleece investors.

Allied Electronics still seems to have some in stock : http://www.alliedelec.com/search/results.aspx?term=silonex,optocoupler
You can find Data Sheets attached to most of Allied's product pages for the Silonex couplers. The Silonex web page used to have a very good technical section - including, as I recall, reference designs, SPICE models, etc. Perhaps you can find some of these documents stashed on other folks' web sites.

Dale
 
ssm2402.jpg
This device (2 pole) and SSM2404 (4 pole) feature ramped break before make operation...iow clickless switching.
Extensively used in audio broadcasting (radio) and pro-audio worlds.
These are unfortunately nowadays obsolete, but eBay and Google show current sources.
ADI have extensive range of analog switches and crosspoint switches also.

Dan.
 
Back to that ADG1612 Analog Device chip I mentioned, they also have a related intem, the ADG1613 (ADG1612 datasheet and product info | 1 ? Typical On Resistance, ±5 V, +12 V, +5 V, and +3.3 V Quad SPST Switches | Analog Switches | Analog Devices). The 1613 is a 2 gate "make before break". Mouser has the whole family (adg1611,12, and 13) and they are reasonably priced. Certainly better priced than any resistive opto isolators I've seen.

But the point of my thread is to hear opinions as to whether or not digial analog switch chips (any kind) would work as well as a resistive opto isolator, and maybe what drawbacks and advantages each has. because when it comes to audio, I can think of many of occasions where even after looking at all the numerical specs, the part that performed best and worst were a total surprise.
 
Organs of a certain era used a lot of CD4016, CD4066, and CD4052 analog switches to do the keying of the notes. The actual keys switched DC voltage, which controlled the analog switch. The elimination of AC keying occured at the same time as the elimination of palladium/rhodium plating on the key contacts in the popularly priced organs.
Traditional organ sounds stop at about 7000 hz so they were able to use heavy low pass filtering to eliminate switch pops. As these weren't zero crossing switches, the pops were pretty radical if the note switched on at a waveform peak. IMHO the whole analog switch product line variations are variations on these pioneering products. The part number prefix now is MC14xxx or others, RCA the originator has been shut down.
The slowness of the CDS cell to turn on and off seems to be a tremendous advantage in the strictly analog world. At $5 each they had better be superior, and I can't say much for the PC board layout compatibility of the packaging of the Vactrol either .
As far as the human interface of analog switches, the actual copper switches we humans use to control them have "contact" bounce as a common feature, which goes on for some milliseconds, which is very audible. To keep this "keyclick" from being audible, either heavy low cut filtering has to be used, or some sort of debounce circuit has to be used that switches only ones. Once debounced, the analog switch will only switch once, which gets the pop up into the megahertz realm. With a switch going back and forth several times over 10 millisecons, the pop would be down in the thousand hertz range which is right in the middle of the music you are trying to pass.
One form of debounce is the flip flop, that switches once, instead of going back and forth several times. These work best with a double throw copper switch.
 
Last edited:
I have tinkered with both. The Vactrol type devices function as a variable resistor. Some can be found with ratings high enough so they can be directly connected into tube electronics and used to vary the parameters of the tube. Try that with a silicon chip.

Most of the analog switches, and e-pots on the market today restrict the signal voltage swing to between ground and Vcc which can be very restrictive in some situations...especially in tube circuits. If you use enough series resistance the chip won't blow up when overdriven, but the distortion produced doesn't sound good, and there can be a noticeable recovery delay.

Both have their place in an amplifier depending on what you are doing, and how high the maximum signal level is.

Want to play with Vactrols cheap, try these:

Electronic Goldmine - SALE! (Pkg of 50) Audio Opto Coupler
 
All analog switch products suffer problems up to and including blowing up when you take input signal beyond supply rails. They can also latch ON.
The Analog Devices "High Voltage" parts will pass/switch about 35 V pk-pk signals maximum.

The old H11F "Photo Fet" Optocoupler can handle about about 60V pk-pk.

The IRF Solid State relays can handle 200V pk-pk.
PVA1354 (IRF) - Power ic Relay, Optocouplers

If using JFET switches, Analog Switches (CMOS or JFET) or "PhotFET Opto Couplers then you have to be very careful about where in the circuit you use them to ensure you do not exceed those signal swing limits.

Sometimes you just have to use a garden variety mechanical type relay.

Nothing worse than a 2 channel amp using JFET switches to channels select, than when you start getting weird distorted breakthrough from the supposedly OFF Channel. You only ever design one like that and I've done mine.

Cheers,
Ian
 
Voltwide,
That is the absolute best way to go about using jfet or cmos analog switches. That is, use them at a virtual earth point where there is effectively no signal voltage just signal current.
It doesn't help when overdriving results in a gain of zero and the vitual earth point "disappears". The voltages will be clipped by previous SS stage to be within the jfets voltage range so I don't think your back to back diodes do anything except form a low pass filter from that 100K + the diode capacitance.

This scheme is not that suitable for tube amps. First we don't like feedback circuits within the amp, they tend to react badly to overdrive. The '170 is a 40V rated part limiting its use to where you can gurantee that the signal voltage is less than about 36V pk-pk (Actually 40V - the Vgs turn off Voltage).

Cheers,
Ian
 
Last edited:
Disabled Account
Joined 2010
It doesn't help when overdriving results in a gain of zero and the vitual earth point "disappears". The voltages will be clipped by previous SS stage to be within the jfets voltage range so I don't think your back to back diodes do anything except form a low pass filter from that 100K + the diode capacitance.

Ian

to be honest, I do not have a clue what you are talking about...

From my point of view during the off-state, the diodes clamp the signal to +-0.7V and the fet disconnects this from the summing input.
 
Sorry I was'nt clear.

If using an OPAMP then the circuit is fine, because there is no signal voltage at that summing point (summing point = virtual earth).

You do need to watch the gain of the opamp which you have set to 1/10th. 10K/100K = 1/10.
If gain is sufficiently high that the output of the opamp saturates (hits the rail voltage) then the summing point stops working like a summing point (because the incremental gain has gone to 0, that is, more input results in 0 more output).

With +/- 15V rails on the opamp then this circuit should be able to cope with +/- 150V input. The ONLY negative? is that gain of 1/10th.

The same thing can happen if using a tube stage as the active gain element rather than an opamp.
You must set the gain such that the stage never saturates.

I was wrong about the diodes - because I was thinking of the JFEt as a shunt switch to 0V rather than a series switch. Sorry.

Hope this make it clearer.

Cheers,
Ian
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.