Hello all,
some respiratory virus totally knocked me off for a week...Glad thread is gaining interest.
I got the amp on today and there was a good amount of shuffling hum coming out. The volume did affect it so I initially thought I shouldput it through a scope.
After 6 screws to the top, I was amazed to find the metal chassis was indeed U shaped with no cover or sides at all, mounted directly to the wooden top of the cab. That immediately made me think I got rf (sounded exactly like AM) because Marshall didn`t want to spend an extra 45c to shield the amp properly. Went to a friend who built a nice metal U shaped top in like 10 min so solved.
Then I decided to have a look inside. Well, I haven`t ever bought anything so shitty. Sorry for language but I cannot come up with another word.
The MG50CFX is not the same inside as the MG50DFX whose schematic I had (and was posted here as well). The PS section has two caps of 2200uF 50V each on the TDA7932 unit made by a manufacturer called Samyoung. The opamp PS unit had two 2200uF 25V caps made by Nover. I have seen Nover in Arcam amps and receivers. That`s it.
Inside there is a good transformer which is far away, as Fahey already noted, from the circuit boards and sensitive stages. The boards were populated by a horde of more Samyoung cap and most of the coupling caps (10uF) were Nover. The TDA7932 section had caps from another spicy manufacturer - KELNA and the font was ELNA`s but it was definitely KELNA. Really, Marshall?
I threw all away and replaced them with whatever I had on hand but the PS section got 4700uF Vishay and Rybucons, the rest were mostly Nichicon KW which I like a lot and Panasonic FR which I also like a lot
All cabling inside, some pretty long, was ordinary with zero shielding. I have not reassembled the amp but bet it was rf going into the high gain stages. Shielded all with aluminium tape and grounded the shields.
Also discovered a cracked (cold) joint and one that was nto soldered at all (bulb joint as I call it, so solder sitting on top in a bulb shape). The solder quality is absolutely ridicilous, even vintage Creek 4040 amps look better.
The cab is made of large particle size board with bad build quality, one speaker screw was not tightened well and the backboard is painted, well, I paint better in my basement with a car shop spray paint bottle...but on the outside looks shiny bearing the big Marshall logo. I would never ever put my name on such crap even if I don`t respect myself that much.
Opamps are 4558 but have to take out some resistors to see what gain they got. The DFX series schematic hows they have around 500 and I plan on lifting this to about 1000.
Soo, instead of adding a tube stage I think I will just use it for some time and get an all valve amp. I do like the design of this one but its a massive box of crap.
some respiratory virus totally knocked me off for a week...Glad thread is gaining interest.
I got the amp on today and there was a good amount of shuffling hum coming out. The volume did affect it so I initially thought I shouldput it through a scope.
After 6 screws to the top, I was amazed to find the metal chassis was indeed U shaped with no cover or sides at all, mounted directly to the wooden top of the cab. That immediately made me think I got rf (sounded exactly like AM) because Marshall didn`t want to spend an extra 45c to shield the amp properly. Went to a friend who built a nice metal U shaped top in like 10 min so solved.
Then I decided to have a look inside. Well, I haven`t ever bought anything so shitty. Sorry for language but I cannot come up with another word.
The MG50CFX is not the same inside as the MG50DFX whose schematic I had (and was posted here as well). The PS section has two caps of 2200uF 50V each on the TDA7932 unit made by a manufacturer called Samyoung. The opamp PS unit had two 2200uF 25V caps made by Nover. I have seen Nover in Arcam amps and receivers. That`s it.
Inside there is a good transformer which is far away, as Fahey already noted, from the circuit boards and sensitive stages. The boards were populated by a horde of more Samyoung cap and most of the coupling caps (10uF) were Nover. The TDA7932 section had caps from another spicy manufacturer - KELNA and the font was ELNA`s but it was definitely KELNA. Really, Marshall?
I threw all away and replaced them with whatever I had on hand but the PS section got 4700uF Vishay and Rybucons, the rest were mostly Nichicon KW which I like a lot and Panasonic FR which I also like a lot
All cabling inside, some pretty long, was ordinary with zero shielding. I have not reassembled the amp but bet it was rf going into the high gain stages. Shielded all with aluminium tape and grounded the shields.Also discovered a cracked (cold) joint and one that was nto soldered at all (bulb joint as I call it, so solder sitting on top in a bulb shape). The solder quality is absolutely ridicilous, even vintage Creek 4040 amps look better.
The cab is made of large particle size board with bad build quality, one speaker screw was not tightened well and the backboard is painted, well, I paint better in my basement with a car shop spray paint bottle...but on the outside looks shiny bearing the big Marshall logo. I would never ever put my name on such crap even if I don`t respect myself that much.
Opamps are 4558 but have to take out some resistors to see what gain they got. The DFX series schematic hows they have around 500 and I plan on lifting this to about 1000.
Soo, instead of adding a tube stage I think I will just use it for some time and get an all valve amp. I do like the design of this one but its a massive box of crap.
I was thinking about tweaking these exactly as well as possibly replacing the VR parallel to R7. But from what I have seen so far, it is possible the CFX schematic to be different to the DFX series one - the PS section is not the same, already verified.
Hi, do you mean R19 (4K7) to be replaced by a 3K3 in series with a pair of 1N5711 diodes as in the Tbird? R39 on the DFX schematic I have appears to be in the headphone amp section.
I guess C19 should be ommited?
Yes, sorry, my bad eyesight.
Not so much - the 3K3 + diodes is doing much of what Led3/Led4 are doing, just in "tube screamer" mode (but they could do with some R in series to de-fizz them a bit)
You want the bit between U1b and U2a : that is replace C21/R19 with the whole network of (10uf +3K3+15K+300+1N5711 + 15K+1N4148+10uf). Due to the asymmetric clipping the DC point will drift and change the clipping behaviour. Very much part of the "secret sauce" of this design. One of the reasons I will bother building/buying this pedal (along with a micro JCM800 clone)
My only concern looking at the reset of the schematic posted by Printer2 is that you may not have enough V+ and V- to get this to work well. I see one pin says +5V. You need at least 12V rail to rail for the Thunderbird, or so the designers claim
No, it needs to stay.
Last edited:
The caps on the opamps were rated at 35VDC by some reason, but I didn`t measure voltage there. I may put a voltage doubler with a mosfet to powere the chip, or may use a resistor network to drop voltage from the TDA7293 which gets +/-38VDC. Current should be small so a good power rating resistor should do without exploding. Will check out how much this would draw.
I managed to finish with this. What it costed:
Stage 1
- all caps replaced with Nichicon KW
- power supply caps increased to 6800uF and 4700uF (from 2200uF and 1000uF) upgraded to Vishay BC
- fitted regulators with heatsinks
- resoldered cold solder joints and two bulb joints (amp had a really poor job at soldering)
- replaced all pots with new ones (old ones had their shafts loose)
- all internal cabling shielded with tape and shileds to chassis ground
- transformer shielded
- increased gain on the overdrive stage by tweaking two resistor values and a trimmer pot
- fitted a big heatsink (leftover) and welded it to the original one so fan is still in and working (it went on just once when I hit it really hard, otherwise cooling is passive so no noise)
TOTAL COST: EUR23
HOURS: 2.5
Stage 2
- replaced original Chunil corp rubbish with a Celestion G12T75 - I bought the Celestion on an ebay auction for EUR41 including delivery and its UK production (2008) with very few hours in it and then sold the Chunil corp one for EUR25 on a local craigslist
- added polyfill to the backwall of the cab and three of the sides
TOTAL COST: EUR18 (EUR16 for the speaker difference and EUR2 for polyfill and some glue to keep it to the wall)
HOURS: 1
Amp now has very little RF and it comes entirely from the 6m lead I use. When removed, it is silent. Overdrive has some hiss but pretty acceptable. Sounds great and I bet with hte new heatsink will keep up the TDA7923 from thermal issues. I got it for like EUR60 delivered, spent an extra EUR41 to upgrade it, so for EUR101 I would say its pretty decent sounding now. Biggest change came from the speaker replacement, stock Chunil Corp speaker is sh*t.
Stage 1
- all caps replaced with Nichicon KW
- power supply caps increased to 6800uF and 4700uF (from 2200uF and 1000uF) upgraded to Vishay BC
- fitted regulators with heatsinks
- resoldered cold solder joints and two bulb joints (amp had a really poor job at soldering)
- replaced all pots with new ones (old ones had their shafts loose)
- all internal cabling shielded with tape and shileds to chassis ground
- transformer shielded
- increased gain on the overdrive stage by tweaking two resistor values and a trimmer pot
- fitted a big heatsink (leftover) and welded it to the original one so fan is still in and working (it went on just once when I hit it really hard, otherwise cooling is passive so no noise)
TOTAL COST: EUR23
HOURS: 2.5
Stage 2
- replaced original Chunil corp rubbish with a Celestion G12T75 - I bought the Celestion on an ebay auction for EUR41 including delivery and its UK production (2008) with very few hours in it and then sold the Chunil corp one for EUR25 on a local craigslist
- added polyfill to the backwall of the cab and three of the sides
TOTAL COST: EUR18 (EUR16 for the speaker difference and EUR2 for polyfill and some glue to keep it to the wall)
HOURS: 1
Amp now has very little RF and it comes entirely from the 6m lead I use. When removed, it is silent. Overdrive has some hiss but pretty acceptable. Sounds great and I bet with hte new heatsink will keep up the TDA7923 from thermal issues. I got it for like EUR60 delivered, spent an extra EUR41 to upgrade it, so for EUR101 I would say its pretty decent sounding now. Biggest change came from the speaker replacement, stock Chunil Corp speaker is sh*t.
I have used valves as front 4ends but they just sounded clean on their own.
They need to be over driven to get a heavier sound.
I came across an op amp circuit in 1982 that promised to give over driven valve type distortion and it worked ok.
It was a bit dull at first but I found with some extra gain and top end lift it had good bite.
softlimiter | harrabylad | Flickr
Appears most of the distortion actually comes from the output valves. I went to a local shop and had a few pedals demoed - my opinion of course, but I think they do very little to the sound and not worth the cost. After listening to some valve amps I will definitely go this way, I like the Jet City JCA50 and build my own cab.
The Marshall actually turned into a pretty nice amp, I am very happy with it now. With dual humbuckers in my ESP LTD and setting it to 10/12/10 o`clock (bass/mid/treble) I got absolutely brutal distortion out of it and its loud and clean. The G12T75 appears to have V-shaped response in the midrange but shreds really nicely.
The Marshall actually turned into a pretty nice amp, I am very happy with it now. With dual humbuckers in my ESP LTD and setting it to 10/12/10 o`clock (bass/mid/treble) I got absolutely brutal distortion out of it and its loud and clean. The G12T75 appears to have V-shaped response in the midrange but shreds really nicely.
The most important single mod you did was replacing cheesy Korean/Vietnamese/Chinese? generic speaker with a real G12T75 speaker.
Second best was upgrading main filter caps *value* (brand is irrelevant) to proper filtering size.
Playing loud enough to saturate power amp introduces supply ripple *straight* into your speaker, period.
Everything else is fine but just "the icing on the cake" and I am glad you are now happy with your amp sound
Now we deserve a couple YT videos or at least some MP3s showing how you advance on your Guitar playing (and your new improved sound)
Second best was upgrading main filter caps *value* (brand is irrelevant) to proper filtering size.
Playing loud enough to saturate power amp introduces supply ripple *straight* into your speaker, period.
Everything else is fine but just "the icing on the cake" and I am glad you are now happy with your amp sound
Now we deserve a couple YT videos or at least some MP3s showing how you advance on your Guitar playing
(and your new improved sound) - Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.