|yldouright||5/24/2018 2:32 PM|
|Crate GX-40C weirdness|
This one stumped me for weeks so I'd like to know what it could've been. The amp came in for scratchy pots and cutting out. It was dirty and abused so I gave it a good once over but then I discovered the following weirdness.
Symptoms: the amp was fully functional on the headphone output but the chorus circuit didn't output to the speakers. I thought it might be a diode or logic gate that might be keeping the chorus effect from mixing properly but why would it work properly on the headphone/line out? I have the schematic for reference if anyone wants to provide some hints.
|The Dude||5/24/2018 4:57 PM|
|That unit has 2 power amps and runs a true stereo chorus. If the speakers are wired out of phase, the chorus effect is minimal. I would check to see that the speakers are wired correctly. If the headphone out works as it should, I can't think of another cause. Or, one speaker isn't working at all.|
|yldouright||5/27/2018 10:30 AM|
Happy Memorial Day and thanks for response. What you describe would be more like a muted effect and not one completely missing but still, it's worth a try. I've informed the current owner of the suggestion and will report when I get his result. Anyone else have any ideas?
|Jazz P Bass||5/27/2018 12:03 PM|
|Which GX 40C do you have?|
There are two versions.
The older one uses the speaker outputs for the headphone.
The newer DSP model uses a standalone opamp circuit to drive the headphones.
The power amps are separate & distinct circuits.
|yldouright||5/28/2018 1:43 PM|
|@Jazz P Bass|
I have been notified it has separate speaker terminal outputs in the rear above the 8" drivers, the headphone out is about 6'' away from them and does double duty as the line level output (unamplified). Another curiousity is the missing "bright" switch on the front panel. In its place is a tone shaper circuit in the dirty channel that has both a switch and a pot which accentuates/cuts the mids from its 12 o'clock position when you turn it clockwise/counterclockwise. Switching that circuit off didn't disable the shaper but it reduced its effect. If I recall correctly, this switch was also only fully working on the line out (headphone) but not out the two 8" speakers. The switch off position still allowed the tone shaping but there was no difference in the speaker output when the switch went on. Again, what stumped me is why both reverb channels worked perfectly well even with these anomalies.
|Enzo||5/28/2018 2:36 PM|
|Please do not confuse GX40 and G40 and other variations.|
|Jazz P Bass||5/28/2018 7:45 PM|
Is there more than one GX 40C?
If the OP has the DSP circuit, then what I said is true.
The headphone is opamp driven.
So there may be a chorus output circuit issue.
|yldouright||5/29/2018 10:38 AM|
|@Jazz P Bass|
I can confirm the amp is a GX-40C and not a G40 but I can't say if it's a DSP circuit.
It looks like you were right all along. Owner confirms only one speaker output and switching the speaker wires moves the output to the other speaker so it looks like both speakers work and at least one set of speaker leads are good. If I get the amp back, I'll do a continuity check on the J5 leads (red/black) and if they're good, the amp channel is probably blown. It looks like this amp uses the TDA 2040 op-amp, would I better off changing both op-amps or searching out an original replacement? What's a suitable drop in?
|g1||5/29/2018 12:15 PM|
|op-amps are voltage amps, TDA2040 is a power amp IC. |
I would just go with the original type, if you can find one from a reputable source.
|The Dude||5/29/2018 7:13 PM|
|^^^^^^Agree. These guys have 5 in stock.|
Audio lab of Ga
Edit: If one amp works, no reason to change it's output IC. They are independent amps. Don't fix it if it ain't broke.
|Enzo||5/30/2018 11:18 AM|
|DSP is digital signal processing - the FX unit. Look inside. If the amp has a busy little circuit board with 100-leg ICs on it, that would be the DSP card. . Lacking that you'd have one large board with regular stuff scattered about.|
google Crate DSP board, and there are images of one.
|yldouright||5/30/2018 12:04 PM|
I couldn't find a photo of the DSP but my model appears to be a GX-40C+ which doesn't have the DSP controls on the front panel like the GX-40D. I seem to recall two long boards inside that amp; one for the pre-amp connected with two flat teflon multi-conduit cables to the power amp section. I don't recall seeing a big multileg chip and I'm pretty sure the boards had a fairly unsophisticated layout/mask. I noted Ebay has these TDA2040 chips for about $0.60 each in lots of 5. The link above provided by @The Dude doesn't appear to be different from those but I know that chips get produced with different (shrinking silicon wafer) masks and therefore have different parameters. I'd hate to introduce oscillation by having chips with largely mismatched parasitics/parameters or speed. Also, the physical layout has one conduit pair (J5) very close to the chip mounted on the right and the other conduit pair a full 18" from the other chip. I presume these correspond and are distanced in the layout to create the chorus effect. Is this a plausible presumption?
|Enzo||5/30/2018 2:12 PM|
|OK then, GX40C+ is specific enough to get the right drawings. If we don't have it here, you can get it from Loud technologies (Crate's parent) If you have the proper schematic, please post it here.|
Here is a DSP board for future reference.
Basic older chorus amps used the MN3xxx series ICs, little 8-leg guys.
Chorusing amps have two channels - stereo amp, each speaker individually driven. In dry mode the preamp signal goes to both for a dual mono. In chorus mode, one side remains dry signal, now in just one speaker. The dry signal is also sent through the delay/phase circuit to make chorus and then fed to the second speaker. Now we have the dry in one and chorus in the other to phase back and forth with each other. That is much more satisfying than mixing it all together into one like a chorus pedal in front of the amp.
If one power amp dies or its speaker, you lose something. If you lose the chorusing side, you get dry all the time. If you lose the dry side, you get either dry or the unmixed chorus. But problems can arise in the signal routing back in the preamp.
SO what we need to do is very specifically determine what is happening. Diagnose the two power amps first, then see what is actually missing in the preamp. If it all works in the phones, that does point to the power amp circuits.
|yldouright||5/30/2018 3:16 PM|
|Yes, it had an 8 leg DIP and the MN3xxx series rings a bell for me. The chorus switch and pots had absolutely no effect on the speaker output if I recall correctly so it seems I lost the wet channel. I've been told it was the red and black leads that didn't output. Now, just one clarification; we are excluding the reverb tank when describing this wet/dry side. I say this because the reverb continued to work for both the clean and dirty channel. I'll be getting the amp back this weekend and I'll know more then. Thanks for the explanation. A Loudtech search for GX-40C+ yields this page.|
|The Dude||5/30/2018 3:40 PM|
|yldouright||5/30/2018 4:25 PM|
Without being able to spec all the parameters, it's hard to know what's fake about them. I've been burned by parts from "reputable" vendors on op-amps before (LM3886 specifically). To make matters worse, the schematics call for both the TDA2040 and TDA2050H. The vendor you provided sells TDA2040V, how do I know which is the best suited for the supporting electronics of the circuit? Even if you compare the spec sheets, we don't know how close they match the spec because of what I mentioned before about more modern wafer masks.
|The Dude||5/30/2018 4:31 PM|
|I've bought from Audio Lab of Georgia. They are a reputable seller and do not sell fakes. If you'd like to try the eBay IC's, go for it.|
"To make matters worse, the schematics call for both the TDA2040 and TDA2050H. The vendor you provided sells TDA2040V, how do I know which is the best suited for the supporting electronics of the circuit?"
You'll want the 2 sides of the amp to match, so it's simple. Just look at what's in there now and order one of those (if that is indeed even the problem).
|Enzo||5/30/2018 5:54 PM|
|Wet side refers to the side that gets the variable phased signal from the chorus, not the reverb.|
If it seems you have a bad power amp, then find the input to the power amp stage and apply a signal. ANything pass? TDAxxxx power amps are as simple as it gets - most of it is the IC.
Does anyone have the GX40C+ drawing? Mine are in a storage unit across town.
|The Dude||5/30/2018 5:57 PM|
|There's a schematic link in post #14. Here's a direct link for the download page.|
|Enzo||5/30/2018 6:19 PM|
|Oh, well there it is, your headphones are driven from the main signal, so it works. Plugging into the phones jack, activates the four power amp mute JFETs, Q12,13,14,15. One of those transistors fails and you get no sound there.|
But look even simpler - the speaker connections. The speakers in the amp are wired in SERIES with the external speaker jacks. A dirty cutout contact on an external jack will turn off your speaker. Leave the speakers connected and plug an external speaker into the extension jack. Wakes up?
Is the speaker OK?
Other than that, the TDA chip is the whole power amp.
|nosaj||5/30/2018 6:27 PM|
|g1||5/30/2018 6:55 PM|
|The concern with the chinese counterfeits is usually when the chip or transistor is obsolete. Why would there all of a sudden be stockpiles of old stock in China? Or new stock when the labelled manufacturer is not making them anymore. |
Usually they are on ebay and shipping from China. Those 2 factors for any obsolete components are a big red flag.
We call them counterfeit, but often they are just a lower spec part that will sometimes work, but usually blow when asked to do it's full potential.
For example, a 2030 might be re-labelled as a 2050. It's an actual working part, but not what you think it is. For some people it would work, for some it will blow at higher volumes.
A transistor that is supposed to be a 10amp type might only be 2amps.
Just examples but that is the general idea.
|Jazz P Bass||5/31/2018 12:19 AM|
|The LM1875 will work just fine.|
|yldouright||5/31/2018 2:42 PM|
|@Jazz P Bass|
Assuming the pinout is identical, the LM1875 is a much faster chip=more parasitic effect and potential oscillation. This is also a concern when you shrink the same chip to a smaller die.
Agree, out of spec, overspec and differing parasitic parameters all have the potential to create havoc in the original circuit.
I promise to report the result when I get the amp back.
Originally Posted by The Dude
|Enzo||5/31/2018 4:04 PM|
|Don't overthink this. LM1875 is a common 5-leg power amp IC used in the exact same applications as the TDA20xx series.|
The circuit in the amp is simple and right out of the data sheets for the ICs.
The V and H versions of the IC are nothing more than how the legs are pre-bent. The V version is for mounting sticking up so the rear legs stick straight down and the front ones are kinked forwards. The H version is for mounting the IC flat on a board. SO the legs all stick out then band back to poke down into the board holes. It is the exact same part, you can unbend the legs, but I rarely see the H version anyway.
A lot of transistors come in pre-bent form. Like small TO92s. You can get then with three in a flat row - most common - or with the center leg kicked forward, or with it kicked back. All depends on your board layout. The parts are the same though.
You want the two amps in your unit to be the same. They don't need to be matched whatever that means here. You just don't want a far flung audible difference. SO just the same type number is good enough. This is just a guitar amp, the circuits are very stable by design. And the freq capability of the IC won't matter much to a circuit that rolls off by 5kHz into speakers with even less range.
|yldouright||6/1/2018 8:58 AM|
Thanks for clearing that up, I thought the H/V designation could have meant they were layed out differently in the package. In your advice, it would seem a TDA2050H or TDA2051H would also be suitable replacements for the TDA2040 that could be in there now. Again, I'll know more tomorrow but why wouldn't changing both chips to the same lot be prefereable to seeking out an obsolete chip to match?
|Enzo||6/1/2018 11:59 AM|
|For the cheap price of two chips, I don't care which you use. I just wanted to point out that having both ICs from the same batch or something is pointless precision. If they are both TDA2050 that is close enough.|
How are your existing ICs mounted? Flat against the pc board? Or sticking up vertically from the board?
|yldouright||6/3/2018 11:43 AM|
I ordered 10 TDA2050's from China. If I ever mod this circuit for more power with a higher voltage trafo, the 2050 has more headroom than the 2040. I plan to replace both chips for consistency. The chips lie flat on the heatsink so I'll have to bend the leads to mimic the "H" version. It's confirmed now, the right channel doesn't work. Inserting an exteral speaker on the output jacks shows the left channel is operational and the right channel is not. By the way, the internal speaker continues to work when the external jack is plugged in on the left channel. I expect this is also true of the right channel. Both sets of leads from the board to the speakers are fine. I note evidence of dried liquid in the amp so even though it's unusual for an amp chip to die, my money is on a dead chip and not the JFETS but we'll see when I get the chips.
|J M Fahey||6/3/2018 4:31 PM|
|Wish you luck |
That said, you might find this test report video interesting:
|yldouright||6/12/2018 1:47 PM|
|I received the Ebay TDA2050 chips I ordered and unlike the photo on the listing, they're similar to the ones described in @J M Fahey's post above with the indents in the center. I'm going to test a few in a jig before I put them in the amp. I can't say I wasn't warned|
|J M Fahey||6/12/2018 2:02 PM|
You can bolt them to any heatsink you have available or even a piece of aluminum railing, always with mica and grease (after 2 or 3 it ill be a mess but easy to clean with paper towels), solder 5 wires from board to ICs and drive them into clipping, into a load.
Just donīt risk your speakers, use any old klunker you have around or an 8 or 4 ohm (whatever is approppriate) 25W or more resistor, IC will be fed by Crate power supply, and check whether IC can put out rated 20W (or thereabouts) or puny 12/13W which shows its fakeness.
If you still have a good original TDA measure what *it* puts out, so as not to have unreal expectations.
*Maybe* original just puts out, say, 14 or 15W *by design* so why expect more from the replacement?
Please post results, we are also curious what is really out in the wild.
And if they turn up being "good" ... cool, we take note of supplier.
Price is good, doubt is on performance.