I've been "collecting" parts over the years and decided to use some up to build a "junk parts" gainclone. I've got LM4780 boards from .... I forget! They are either from Peter Daniel or Brian GT? Anyway, I've also got some really neat monoblock chassis cases, a few transformers, XLR connectors, speaker binding posts, sufficient parts for 'parallel' LM4780s, AC inlets and mostly everything you need for a gainclone. This will be 'junk parts' wherever possible, and will be "bridge-parallel" gainclone with XLR inputs. So lets see what kind of Junk i'm using (its really not that junky when you clean the stuff up-- I say junk cause most of this has been sitting around for years.)
Here is the chassis, or at least one of them.
The chassis for these are a pair of neat little heat sink boxes of some kind that I found at the surplus store years ago. They were all grungy but some sanding cleaned them right up. You see I have already started mounting stuff in there, mainly and testing screw hole locations. A hole in the faceplate for a Red overtemperature LED is also drilled.
Overtemperature? This will be bridge-parallel and will probably run HOT. These gainclones run hotter than people think, this will be a little brick heater, hopefully it won't get painfully hot, but the case is only 10 inches long by 4 inches wide by 3 inches tall, very small for what might be a 200W amp. Yes, am worried about heat.
The holes will be covered by a metallic grill so screws and holes will not be visible. I also drilled some extra holes in the cover so that this does not turn into a little oven for the components inside.
Here is the chassis, or at least one of them.
An externally hosted image should be here but it was not working when we last tested it.
The chassis for these are a pair of neat little heat sink boxes of some kind that I found at the surplus store years ago. They were all grungy but some sanding cleaned them right up. You see I have already started mounting stuff in there, mainly and testing screw hole locations. A hole in the faceplate for a Red overtemperature LED is also drilled.
Overtemperature? This will be bridge-parallel and will probably run HOT. These gainclones run hotter than people think, this will be a little brick heater, hopefully it won't get painfully hot, but the case is only 10 inches long by 4 inches wide by 3 inches tall, very small for what might be a 200W amp. Yes, am worried about heat.
The holes will be covered by a metallic grill so screws and holes will not be visible. I also drilled some extra holes in the cover so that this does not turn into a little oven for the components inside.
An externally hosted image should be here but it was not working when we last tested it.
Back Panel
This will house two LM4780s, each in "parallel" mode, and each running out of phase with the other, to be a "bridge parallel" amplifier. People here call this a "BPA 200" I think for 'bridge parallel 200 WPC'?? Not going to try to make this 200 WPC because I've got small cases but certainly this will be more powerful than your standard chipamp.
So I have these two XLR inputs that have been sitting around for years, this amp will accept an XLR input, DC input, and have a speaker output. The amp is not intended to work signals that are not differential.
The + signal from the XLR goes to one LM4780, the - signal to another, and the output is taken across them to the speaker terminals.
The DC signal will be carried by a set of IEC 319/320 connectors and cables. Looks like IEC319/320 carries 220 V at over 15 amps, so should be more than sufficient. But the connectors are big. I just have enough room for them. I did buy these new as I had no parts in the junk box for making DC connections. These were far cheaper than any Neutrik part and cables and connectors are very plentiful.
So on we go-- how do you make large holes in a metal panel? With a hand drill and a file.
After drilling a 1/4" hole, pull out the file and go to town. If you are careful (outline the desired opening), it takes about 15 minutes to rout/ file a huge hole. No fancy punches needed. Looks crappy but just wash, sand, wash.
This will house two LM4780s, each in "parallel" mode, and each running out of phase with the other, to be a "bridge parallel" amplifier. People here call this a "BPA 200" I think for 'bridge parallel 200 WPC'?? Not going to try to make this 200 WPC because I've got small cases but certainly this will be more powerful than your standard chipamp.
So I have these two XLR inputs that have been sitting around for years, this amp will accept an XLR input, DC input, and have a speaker output. The amp is not intended to work signals that are not differential.
The + signal from the XLR goes to one LM4780, the - signal to another, and the output is taken across them to the speaker terminals.
The DC signal will be carried by a set of IEC 319/320 connectors and cables. Looks like IEC319/320 carries 220 V at over 15 amps, so should be more than sufficient. But the connectors are big. I just have enough room for them. I did buy these new as I had no parts in the junk box for making DC connections. These were far cheaper than any Neutrik part and cables and connectors are very plentiful.
So on we go-- how do you make large holes in a metal panel? With a hand drill and a file.
An externally hosted image should be here but it was not working when we last tested it.
After drilling a 1/4" hole, pull out the file and go to town. If you are careful (outline the desired opening), it takes about 15 minutes to rout/ file a huge hole. No fancy punches needed. Looks crappy but just wash, sand, wash.
An externally hosted image should be here but it was not working when we last tested it.
Mounting the Chips
In go the boards and chips.
Here is the Schematic I used- Bridge Parallel XLR Input Schematic.
EDIT- as you see this must be driven from a fully balanced source, and is a noninverting setup. Also, I have made absolutely no provision to reduce DC offset or incorporate any protection of any kind. Wish me luck (hey, I'm not that smart and this is a junk parts project).
Here is one of the mono blocks with some of the components installed.
I had some issues- the 2,200 uF capacitors I had were snap in and did not have the right pitch for the boards, so I had to drill out some holes. The current sharing resistors I had were .1 ohm, but it is better and safer to use .2 ohm. I used .1 ohm anyway.
I did not have enough .1 uF fancy Wima caps to bypass the electrolytics so I used the caps I had, mounted on the underside of the electrolytics. I think they are stacked film?
The chips are mounted to the roof of the case which is pretty solid. I used two standoffs to connect one edge of the PCB to the side of the case for extra stability. I am not bothering with two standoffs. One required enough precision to make me nervous that I could pull it off. Yes, I had some very long screws around to use with the standoffs; filled up a bag of various screws at the surplus store many years ago and have a lot left.
A Stancor bimetal switch is also mounted to the roof of the heat sink box, this will close at (I think) about 63 degrees C and connect the + supply rail to a red LED through a current limiting resistor, which indicates that the case is really hot. Had a bunch of these temperature switches left over from previous projects, I tend to use them a lot. I am going to have power LEDs in the separate power supply so these little boxes will run with no illumination other than the overtemp LED.
In go the boards and chips.
Here is the Schematic I used- Bridge Parallel XLR Input Schematic.
EDIT- as you see this must be driven from a fully balanced source, and is a noninverting setup. Also, I have made absolutely no provision to reduce DC offset or incorporate any protection of any kind. Wish me luck (hey, I'm not that smart and this is a junk parts project).
Here is one of the mono blocks with some of the components installed.
An externally hosted image should be here but it was not working when we last tested it.
I had some issues- the 2,200 uF capacitors I had were snap in and did not have the right pitch for the boards, so I had to drill out some holes. The current sharing resistors I had were .1 ohm, but it is better and safer to use .2 ohm. I used .1 ohm anyway.
I did not have enough .1 uF fancy Wima caps to bypass the electrolytics so I used the caps I had, mounted on the underside of the electrolytics. I think they are stacked film?
The chips are mounted to the roof of the case which is pretty solid. I used two standoffs to connect one edge of the PCB to the side of the case for extra stability. I am not bothering with two standoffs. One required enough precision to make me nervous that I could pull it off. Yes, I had some very long screws around to use with the standoffs; filled up a bag of various screws at the surplus store many years ago and have a lot left.
A Stancor bimetal switch is also mounted to the roof of the heat sink box, this will close at (I think) about 63 degrees C and connect the + supply rail to a red LED through a current limiting resistor, which indicates that the case is really hot. Had a bunch of these temperature switches left over from previous projects, I tend to use them a lot. I am going to have power LEDs in the separate power supply so these little boxes will run with no illumination other than the overtemp LED.
Last edited:
Hi,
BPA using a pair of 4780 sounds nice.
Most build a 4780 as a parallel. This draws double the current through the supply pins.
I think that building the 4780 as bridged might be better since the currents run into and out of the supply pins is not doubled. (double the power into double the impedance is equal to saying @ the same current as the single).
Then a pair of these 4780 to give your parallel combination.
I don't know if this would be measurable nor even audible but worth trying.
Your XLR feed. Is that a balanced hot and cold signal pair?
Are you going to use the chips as inverting or all as non-inverting?
Are you planning on using DC servos to balance the output voltages?
BPA using a pair of 4780 sounds nice.
Most build a 4780 as a parallel. This draws double the current through the supply pins.
I think that building the 4780 as bridged might be better since the currents run into and out of the supply pins is not doubled. (double the power into double the impedance is equal to saying @ the same current as the single).
Then a pair of these 4780 to give your parallel combination.
I don't know if this would be measurable nor even audible but worth trying.
Your XLR feed. Is that a balanced hot and cold signal pair?
Are you going to use the chips as inverting or all as non-inverting?
Are you planning on using DC servos to balance the output voltages?
Andrew, yes I thought of this but dismissed it thinking it wouldn't work. Had you provided this suggestion before I populated the boards I would have had the incentive to try it out. (Of course you did not know that I was about to build this so could not suggest anything). Another good idea from AndrewT, but its too late now.
Last edited:
Wiring the Monoblocks
I bought a bunch of teflon insulated silver coated wire a few years ago. Its not the fattest wire out there but it is very dense and sounds fantastic. Just kidding, I have no idea how it 'sounds', but its pretty nice wire, so I'm using it to wire the AC and DC connections in the monoblocks.
Here are the monoblocks wired up.
There is a CL-60 thermister from DC ground (PS ground) to the case in order to ground but decouple the case.
For the XLR input, there is only one ground lead but I need two, one to go to each board. So I just jumpered the signal ground from one board to the other (yellow striped wire).
I bought a bunch of teflon insulated silver coated wire a few years ago. Its not the fattest wire out there but it is very dense and sounds fantastic. Just kidding, I have no idea how it 'sounds', but its pretty nice wire, so I'm using it to wire the AC and DC connections in the monoblocks.
Here are the monoblocks wired up.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
There is a CL-60 thermister from DC ground (PS ground) to the case in order to ground but decouple the case.
For the XLR input, there is only one ground lead but I need two, one to go to each board. So I just jumpered the signal ground from one board to the other (yellow striped wire).
Screws?
Somewhere along the way I lost some of the short stubby screws that hold in the front and back panels. These were standard countersunk flat head screws.
I don't think they look all that good anyway, so I cut up some phillips head screws using my crimp tool.
These tools are supposed to cut the screw threads cleanly so it goes right into the threaded hole. Of course it really doesn't and is unreliable. You still may have to file or rework the end of the screw where the cut was made. The screws here need to be very short so I was forced to make these cuts or to drill and retap. Got it working after a while.
Gotta say that small things like getting the right look with the screws and fasteners mean a lot.
I think this looks better than the slot head version.
Somewhere along the way I lost some of the short stubby screws that hold in the front and back panels. These were standard countersunk flat head screws.
An externally hosted image should be here but it was not working when we last tested it.
I don't think they look all that good anyway, so I cut up some phillips head screws using my crimp tool.
An externally hosted image should be here but it was not working when we last tested it.
These tools are supposed to cut the screw threads cleanly so it goes right into the threaded hole. Of course it really doesn't and is unreliable. You still may have to file or rework the end of the screw where the cut was made. The screws here need to be very short so I was forced to make these cuts or to drill and retap. Got it working after a while.
An externally hosted image should be here but it was not working when we last tested it.
Gotta say that small things like getting the right look with the screws and fasteners mean a lot.
I think this looks better than the slot head version.
Last edited:
Power Supply
I've got a small, and I mean small, box for the power supply. Its 9" wide by 7" deep by 3" tall. Just barely tall enough for a power transformer or power capacitors. I have to be really careful to get everything to fit.
I have a lot of those 15,000 uF 63V capacitors from Apex Jr., so in they go. I cannot use a PCB because any stand-off will exceed the 3 inches of height that I have. The capacitors have to be hot-glued to the case itself.
The case has grooves for fitting in the front and back panels, they slide into the base and are held in the grooves. The cover also has mating grooves around its periphery to hold the front and back panels.
Its kind of neat to have a case with absolutely no screws holding the front and back panels in. The bottom and top are connected by long screws and the grooves hold the panels in.
I've got a small, and I mean small, box for the power supply. Its 9" wide by 7" deep by 3" tall. Just barely tall enough for a power transformer or power capacitors. I have to be really careful to get everything to fit.
I have a lot of those 15,000 uF 63V capacitors from Apex Jr., so in they go. I cannot use a PCB because any stand-off will exceed the 3 inches of height that I have. The capacitors have to be hot-glued to the case itself.
An externally hosted image should be here but it was not working when we last tested it.
The case has grooves for fitting in the front and back panels, they slide into the base and are held in the grooves. The cover also has mating grooves around its periphery to hold the front and back panels.
Its kind of neat to have a case with absolutely no screws holding the front and back panels in. The bottom and top are connected by long screws and the grooves hold the panels in.
Power Supply Back Panel
I drilled and filed the back panel to fit the connectors. Again, no special tools needed. Then I used some 2-ton epoxy to epoxy small sheet metal cut outs behind the holes that I was not using. Finally I spraypainted the entire thing white. I had 2 cans of white and 2 cans of black, so used some black for the top of the case.
If I had to to it again, I'd spraypaint the metal pieces and back panel separetly BEFORE epoxing everything together. Duh!
Here is how it turned out.
Its ok, but could have been much better.
As you can see the fuseholder in the pics is different- I used a surplus fuse holder at first but it was so fragile (old) that the metal tabs simply broke off. I put some allen head screws in the back panel just to fill up the screw holes; had a lot of allen head screws left over from other projects.
I drilled and filed the back panel to fit the connectors. Again, no special tools needed. Then I used some 2-ton epoxy to epoxy small sheet metal cut outs behind the holes that I was not using. Finally I spraypainted the entire thing white. I had 2 cans of white and 2 cans of black, so used some black for the top of the case.
If I had to to it again, I'd spraypaint the metal pieces and back panel separetly BEFORE epoxing everything together. Duh!
Here is how it turned out.
Its ok, but could have been much better.
As you can see the fuseholder in the pics is different- I used a surplus fuse holder at first but it was so fragile (old) that the metal tabs simply broke off. I put some allen head screws in the back panel just to fill up the screw holes; had a lot of allen head screws left over from other projects.
Last edited:
DC Wiring
I know you are supposed to keep the DC away from the AC, which is hard to implement in such a small case. Here I have wired the DC connectors to the PS caps along the back panel. They will be perpendicular to the AC inlet wires which cross above them. Not a perfect situation since they should be remote from the AC wires, but since they are perpendicular they should not pick up too much of the 60 Hz.
Point to point soldering and hand wiring.
No room to fit in connectors for the bridge and transformer hookups.
The inside of this PS is not going to win any beauty contests.
I know you are supposed to keep the DC away from the AC, which is hard to implement in such a small case. Here I have wired the DC connectors to the PS caps along the back panel. They will be perpendicular to the AC inlet wires which cross above them. Not a perfect situation since they should be remote from the AC wires, but since they are perpendicular they should not pick up too much of the 60 Hz.
An externally hosted image should be here but it was not working when we last tested it.
Point to point soldering and hand wiring.
No room to fit in connectors for the bridge and transformer hookups.
The inside of this PS is not going to win any beauty contests.
Last edited:
Power Supply Front Panel
As I was trying to figure out what to do with the front panel of the power supply I ran across an unused sheet of white acrylic. Lucky me! If I knew I had this I would have used it to block off the unused connections on the BACK panel! Duh!
Luckily I found it in time to use it for the front panel. My plan is to cover the front panel in white acrylic and put power-on LEDs BEHIND the panel, they should shine through the white acrylic since it is partially light transmissive.
I cut the acrylic, clamped it to the front panel, traced where the holes should go, and drilled away. This stuff is much easier to file than metal, whew.
I still had to drill and file the metal panel to fit in the on/off switch.
While the panel has no screws holding it in place, I put in allen head screws that simply hold the acrylic to the metal plate. The acrylic is slightly smaller than the front plate so that the front plate can fit into the gooves on the case. I don't mind if its not an exact fit. This should look ok if it turns out to be centered.
Well the screw holes are not exactly even, but it is a nice smooth panel. The LEDs will go into two metal openings behind the acrylic, and hopefully shine through.
As I was trying to figure out what to do with the front panel of the power supply I ran across an unused sheet of white acrylic. Lucky me! If I knew I had this I would have used it to block off the unused connections on the BACK panel! Duh!
Luckily I found it in time to use it for the front panel. My plan is to cover the front panel in white acrylic and put power-on LEDs BEHIND the panel, they should shine through the white acrylic since it is partially light transmissive.
I cut the acrylic, clamped it to the front panel, traced where the holes should go, and drilled away. This stuff is much easier to file than metal, whew.
An externally hosted image should be here but it was not working when we last tested it.
I still had to drill and file the metal panel to fit in the on/off switch.
An externally hosted image should be here but it was not working when we last tested it.
While the panel has no screws holding it in place, I put in allen head screws that simply hold the acrylic to the metal plate. The acrylic is slightly smaller than the front plate so that the front plate can fit into the gooves on the case. I don't mind if its not an exact fit. This should look ok if it turns out to be centered.
An externally hosted image should be here but it was not working when we last tested it.
Well the screw holes are not exactly even, but it is a nice smooth panel. The LEDs will go into two metal openings behind the acrylic, and hopefully shine through.
Last edited:
this does not sound right.
There is no Signal Ground in a balanced XLR cable/plug.
Pin1 goes to chassis.
Pins2 & 3 go to the balanced impedance differential inputs.
The amplifier (a pair of bridged amps) amplifies the difference between the two Signals. It does not reference Pin1.
These boards are ground happy, they have at least 4 pads for Ground, whihc are Ground (DC+), Ground (DC-), Ground (Input), Ground (output), all of which are connected to the same big ground plane on the board.
The DC supplies connect to +, -, and one of the DC grounds. I have also connected the DC ground to the chassis.
The input ground from the XLR connector goes to the same big ground plane on both boards.
If I were to change this and connect the XLR ground instead right to the case, it would still be connected to the DC power ground, which is connected to the big ground pad on the board.... so the situation is identical, all the connections are effectively the same. The input XLR ground would be shorted to the big ground pad just like it is now.
Except of course the wiring and layout would be different so it would affect ground loops, noise pickup, etc...
Not sure which way to go.
Power Supply
I've got a few transformers, the one I'm using is a 500 VA 20-0-20. It was the only one that would fit in the case, and it barely fits. Also, this should give me DC rails of about 27 volts, slightly lower than the 35 volts that the chipamps are capable of using. Good thing, because these will be getting hot in the bridge (parallel) configuration and a lower voltage will lower the heat.
Here is the completed power supply.
The transformer is wrapped in mu metal shielding, which is itself wrapped in electrical tape so as not to short out the AC input and the on/off switch contacts, which just barely have enough clearance not to touch.
The LEDs are wired in using 3W current limiting resistors. No snubbers, just a bleed resistor across the capacitors.
There is a line rated .047 uF cap across the AC terminals (after the fuse) to reduce AC noise, and the saftey earth is connected to the case on the back panel. The DC ground is connected to the saftey earth through a 5W, 18 ohm resistor to decouple circuit ground from earth ground.
All the parts of the case are painted and I would not suspect that voltage could appear anywhere on the case, or if it did, it may be only on 1 part and not on other parts. So it makes it difficult to connect to saftey earth since there are no good connection points and what are you going to do- connect the front, top, bottom, back panels to each other with wire?
I've got a few transformers, the one I'm using is a 500 VA 20-0-20. It was the only one that would fit in the case, and it barely fits. Also, this should give me DC rails of about 27 volts, slightly lower than the 35 volts that the chipamps are capable of using. Good thing, because these will be getting hot in the bridge (parallel) configuration and a lower voltage will lower the heat.
Here is the completed power supply.
The transformer is wrapped in mu metal shielding, which is itself wrapped in electrical tape so as not to short out the AC input and the on/off switch contacts, which just barely have enough clearance not to touch.
The LEDs are wired in using 3W current limiting resistors. No snubbers, just a bleed resistor across the capacitors.
There is a line rated .047 uF cap across the AC terminals (after the fuse) to reduce AC noise, and the saftey earth is connected to the case on the back panel. The DC ground is connected to the saftey earth through a 5W, 18 ohm resistor to decouple circuit ground from earth ground.
All the parts of the case are painted and I would not suspect that voltage could appear anywhere on the case, or if it did, it may be only on 1 part and not on other parts. So it makes it difficult to connect to saftey earth since there are no good connection points and what are you going to do- connect the front, top, bottom, back panels to each other with wire?
PS again
Here is the PS again. You can see the blue leds shining through the plastic. Its brighter than it looks.
I'm getting 27.9 VDC unloaded.
Here is the PS again. You can see the blue leds shining through the plastic. Its brighter than it looks.
I'm getting 27.9 VDC unloaded.
An externally hosted image should be here but it was not working when we last tested it.
Complete Amps
The amps are complete.
They are quite small.
The amps are complete.
An externally hosted image should be here but it was not working when we last tested it.
They are quite small.
An externally hosted image should be here but it was not working when we last tested it.
Results
Sound- surprisingly these have very deep bass extension, much better than the parallel version or the standard version. They are detailed like you would expect from chip amps. No hum at all despite the grounding issues discussed earlier. With your ear at the tweeters, there is just a very faint hiss of HF noise. I listened to them for about 8 hours so far and was surprised by how good they sounded.
Technical- they do get hot. After 4 hours at medium volume the tops of the little bricks were about 57 degrees C. I have never got them into the 60's and the little red lights have not lit yet. I can stick a probe intot the case through a screw hole, and the inside air was 47C after 4 hours. The plan is to perf the bottom panel to get some airflow in there. I am sure that if I stressed the amps they would get even hotter. But that's what you get when using up all your junk parts- it runs too hot and so is less than ideal. So you have to worry about the temperature protection in the chips kicking in, although I didn't detect anything so far.
DC offset at the output is less than 5mV Open Circ, and goes down with the inputs shorted. This despite the fact that there is nothing included to adjust the offset or reduce it. Not bad.
Sound- surprisingly these have very deep bass extension, much better than the parallel version or the standard version. They are detailed like you would expect from chip amps. No hum at all despite the grounding issues discussed earlier. With your ear at the tweeters, there is just a very faint hiss of HF noise. I listened to them for about 8 hours so far and was surprised by how good they sounded.
Technical- they do get hot. After 4 hours at medium volume the tops of the little bricks were about 57 degrees C. I have never got them into the 60's and the little red lights have not lit yet. I can stick a probe intot the case through a screw hole, and the inside air was 47C after 4 hours. The plan is to perf the bottom panel to get some airflow in there. I am sure that if I stressed the amps they would get even hotter. But that's what you get when using up all your junk parts- it runs too hot and so is less than ideal. So you have to worry about the temperature protection in the chips kicking in, although I didn't detect anything so far.
DC offset at the output is less than 5mV Open Circ, and goes down with the inputs shorted. This despite the fact that there is nothing included to adjust the offset or reduce it. Not bad.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Testing
Tried to get it to clip- here is a 20 volts per division differential measurement at the output into 5 ohms resistive load. At this time I was using a different, smaller (and higher voltage), transformer in the power supply. Just testing to see if the thing worked.
Tried to get it to clip- here is a 20 volts per division differential measurement at the output into 5 ohms resistive load. At this time I was using a different, smaller (and higher voltage), transformer in the power supply. Just testing to see if the thing worked.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I was astonished with while paralleling 3 lm3886's (I built a bpa300.) How close the resistors should be matched for good operation. The output 0.22ohm resistor doesn't really matter that much but the other do a lot!
I did use 1% resistors but did not test them. After about a month of use the one chip went up in smoke and put -35V on the output.
After replacing the broken one I measured the temperature one each chip while putting it under high load just to see one chip shooting up to 55C while the other two stayed at 35C.
I now have expensive 0.1% resistors in my parallel circuit and all stay within 5C from each other even at high load.
oh yes
+10 for the radiohead album lgreen 🙂
I did use 1% resistors but did not test them. After about a month of use the one chip went up in smoke and put -35V on the output.
After replacing the broken one I measured the temperature one each chip while putting it under high load just to see one chip shooting up to 55C while the other two stayed at 35C.
I now have expensive 0.1% resistors in my parallel circuit and all stay within 5C from each other even at high load.
oh yes
+10 for the radiohead album lgreen 🙂
- Status
- Not open for further replies.