I have black anodized screws for the chassis, courtesy of Landfall. I also had a big bag of 6-32 and 8-32 stainless screws from another project so I decided to use them to attach the components to the chassis. So I researched blackening...and here we are. My thought processes tend to be non-linear, maybe it's that scientist vs engineer thing, but I thought it would be cool to blacken them myself, being a chemist and all...
Hey, it's better than goin' all Walter White and Breaking Bad, ya know?
Fair enough. The purpose of a hobby is to spend your money and your time. Blackening screws seemed to fit the bill nicely... 🙂
~Tom
I'll be assembling steadily over the next few weeks. As Tom said, this is a hobby.
I do have 19 vacation days left to take this year, so that should speed things up a bit.
Tonight I assembled one of the chassis. First, I wire brushed all the corner areas where the chassis panels and top connect together to get a good clean aluminum surface and chromated them using an Alodyne Felt Tip applicator. The chromated areas will be corrosion resistant and will remain conductive. This ensures I have a good electrical connection between all of the panels and the safety ground. It also provides some shielding of the circuits. Testing showed continuity and very low resistance between all of the panels.
I attached a snapshot of the assembled, empty chassis. Next step is to install and test the filament circuit.
I attached a snapshot of the assembled, empty chassis. Next step is to install and test the filament circuit.
Attachments
I wired in the filament power supply this weekend, and in the process found out that Hammond 159ZL chokes CAN be used in a 10A choke input power supply if they are modded just a little. Wire used was 10ga stranded copper.
I went for the 159ZL chokes for both positions in the LCLC supply because they are substantial, yet not TOO big and heavy (a 20 amp rated choke of similar inductance is just too large). They are only rated at 10A, 2.5mH, but in speaking with Hammond they told me that they would be OK in a choke input supply from an 18V transformer providing they didn't pass more than 12A rms. According to PSUD my supply hits them with 10.97A rms, so in that respect, according to Hammond, they were fine. Hammond was mostly concerned with core heating, not buzzing.
I was concerned they would buzz since they are only held together with a stamped steel frame. Hammond kindly includes 6 holes around the perimeter of the core, however, and if you're careful, you can drill out the frame without mangling the core and put screws through the frame to clamp it down much tighter. I also used channel lock pliers to bend the frame tightly around the core before tightening down the screws. Picture 1 shows a 159ZL with the screws in place versus one that is stock.
After I assembled it all in the chassis (Picture 2) I added a load to the PS, namely three 4ohm, 100W dummy loads in parallel to make a 1.333ohm load, and brought it all up on a variac. At 15.1V output, which corresponds to 11.3A, all was well, with only the slightest buzzing from L1. The buzzing was only audible with my ear within 3 inches of the choke, not bad at all. Voltage was right where PSUD said it would be, and after running for 30 minutes, the temp of the core of L1 was only 48degC as measured by infrared thermometer.
So, it looks like this PS should be fine for the filament. Bending the choke input rules a bit, but still OK. BTW, the transformer is 20A, 18V.
I went for the 159ZL chokes for both positions in the LCLC supply because they are substantial, yet not TOO big and heavy (a 20 amp rated choke of similar inductance is just too large). They are only rated at 10A, 2.5mH, but in speaking with Hammond they told me that they would be OK in a choke input supply from an 18V transformer providing they didn't pass more than 12A rms. According to PSUD my supply hits them with 10.97A rms, so in that respect, according to Hammond, they were fine. Hammond was mostly concerned with core heating, not buzzing.
I was concerned they would buzz since they are only held together with a stamped steel frame. Hammond kindly includes 6 holes around the perimeter of the core, however, and if you're careful, you can drill out the frame without mangling the core and put screws through the frame to clamp it down much tighter. I also used channel lock pliers to bend the frame tightly around the core before tightening down the screws. Picture 1 shows a 159ZL with the screws in place versus one that is stock.
After I assembled it all in the chassis (Picture 2) I added a load to the PS, namely three 4ohm, 100W dummy loads in parallel to make a 1.333ohm load, and brought it all up on a variac. At 15.1V output, which corresponds to 11.3A, all was well, with only the slightest buzzing from L1. The buzzing was only audible with my ear within 3 inches of the choke, not bad at all. Voltage was right where PSUD said it would be, and after running for 30 minutes, the temp of the core of L1 was only 48degC as measured by infrared thermometer.
So, it looks like this PS should be fine for the filament. Bending the choke input rules a bit, but still OK. BTW, the transformer is 20A, 18V.
Attachments
Just wanted to relate an excellent experience I just had with Lars Clausen of NewclassD. I am using their softstart module on the 833 B+ circuit for a few reasons:
It can handle the load
It has a built-in DC and EMI filter
It has the option of using a low voltage pushbutton to turn it on and off from the front panel, so I can use the mains switch to turn on the driver, bias and filament circuits, then use the pushbutton after a suitable delay to start the 833 B+.
The only issue is that it is set to default to always on, and to remember what it was set at in case of a power failure. So...if there is a power failure, or if the mains power is switched off with the B+ still on, when power is restored the B+ would be instantly on. This is fine for a class D amp, such as it was designed for, but not for this amp!
I sent NewclassD an email about this and after some discussions Lars suggested that he could program a new controller chip and send it to me, with the default changed to OFF. All I need to do is solder it in and all should be good! No charge for the programming or the shipping, either!
Great customer service, Lars!
It can handle the load
It has a built-in DC and EMI filter
It has the option of using a low voltage pushbutton to turn it on and off from the front panel, so I can use the mains switch to turn on the driver, bias and filament circuits, then use the pushbutton after a suitable delay to start the 833 B+.
The only issue is that it is set to default to always on, and to remember what it was set at in case of a power failure. So...if there is a power failure, or if the mains power is switched off with the B+ still on, when power is restored the B+ would be instantly on. This is fine for a class D amp, such as it was designed for, but not for this amp!
I sent NewclassD an email about this and after some discussions Lars suggested that he could program a new controller chip and send it to me, with the default changed to OFF. All I need to do is solder it in and all should be good! No charge for the programming or the shipping, either!
Great customer service, Lars!
Completed the filament supply today by adding the two Coleman regulators in parallel to get 10V@10A. I tested it by hooking up four 100W, 4ohm dummy loads in parallel to get a 1ohm load. 15.0V into the regulators gave 10.0V across the load after adjustment. After 15 minutes of running, the heat sink was only up to 50 degC, and the top lid of the chassis was at 46 degC, so it looks like with the fan running there should be no thermal issues.
Still a slight buzzing from L1, so I mounted it on rubber washers for isolation. Hopefully the mass loading of the lid with >100lbs of iron will be sufficient to curtail any microphonics issues.
BTW, I moved the adjustment pots off the boards and mounted them side by side on the top panel so adjustments can be made without having to open up the amps; at 150lbs each, that will be a back-saver, for sure.
Still a slight buzzing from L1, so I mounted it on rubber washers for isolation. Hopefully the mass loading of the lid with >100lbs of iron will be sufficient to curtail any microphonics issues.
BTW, I moved the adjustment pots off the boards and mounted them side by side on the top panel so adjustments can be made without having to open up the amps; at 150lbs each, that will be a back-saver, for sure.
Attachments
Looks good, Bob! I certainly like the thermal spreader. That will avoid any risk of transient thermal over-stresses - during a warm-start for instance.
Thanks, Rod. Your regulators work perfectly, of course! I used the clips and thermal pad "socks" so there shouldn't be any electrical insulation issues when I float them up to 250V. Thanks to you and Tom for those suggestions.
.
Last edited:
Well done! Interested in a close up picture of the clips used and thermal pads if you have them?
Keen to see your progress on this. I've been using Rod's filament boards with excellent results in the 4-65a SE amp with 46 driver in filament bias. All DHT and DC coupled so thermal stability of the Coleman regulators is a must in my design!
Keen to see your progress on this. I've been using Rod's filament boards with excellent results in the 4-65a SE amp with 46 driver in filament bias. All DHT and DC coupled so thermal stability of the Coleman regulators is a must in my design!
The thermal pads:
http://www.digikey.com/product-sear...=us&keywords=BER156-ND&formaction=on&x=9&y=19
The clips:
TSC607-ZP - LAIRD TECHNOLOGIES - CLIP, 2 X, TO-220 | Newark
The clips are made to hold two TO-220 chips, one on each side, but I just used 1/2 of each one. I was going to snip off the other side but then I figured it might give me a tiny bit more dissipation leaving that little "fin" there. It can't hurt, and it's less work, so...
I'm using the same TO-220 clips on the big TO-247 mosfet in the shunt regulator in the 833 cathode, but I'm using two of them per chip, one on each side of the chip - they fit perfectly that way and will certainly hold the chip FIRMLY to the heat sink. I also purchased TO-247 "socks" for that chip, too.
Last edited:
'm not familiar with the Coleman regulators (I build my own stuff). Are they linear or switchers?
Thanks Magz.
@defiant: Rod's regulators are really good. A refined circuit over the years to provide slow ramp up of current and very low noise filament current. It's based on a gyrator and CCS to isolate filament current from transformer.
Many of us in this forum have used them with great success in multiple designs from preamps (e.g 26, 01a, 10Y) to output stages (e.g. 45,46, 4P1L, 300B, 2a3, 813, gm-70, etc.)
Get in touch with Rod if you need further info.
Cheers, Ale
@defiant: Rod's regulators are really good. A refined circuit over the years to provide slow ramp up of current and very low noise filament current. It's based on a gyrator and CCS to isolate filament current from transformer.
Many of us in this forum have used them with great success in multiple designs from preamps (e.g 26, 01a, 10Y) to output stages (e.g. 45,46, 4P1L, 300B, 2a3, 813, gm-70, etc.)
Get in touch with Rod if you need further info.
Cheers, Ale
I was really asking about dissipation in a roundabout manner. 833 filaments are thirsty, and even with a relatively low voltage across the transistors they will be pouring off the heat. Hopefully the sinks are large enough.
The current monitor resistors get up to 110C or so in my tests, but the heat sink only reached 50C over the course of 15 minutes. I'm using two regulators in parallel at 5A each, so the dissipation is spread across four chips plus the resistors. In any case, there is a fan that sucks warm air from the chassis at up to 33cfm and routes it around the 833, so cooling shouldn't be an issue. At 33cfm, all the air inside the chassis will be exchanged in less than 2 seconds.
Hi Magz,
As hpeter wrote here: http://www.diyaudio.com/forums/tube...e-crisis-my-833c-amp-build-7.html#post3431529
This are 50W, only 😉,...+30W for the choke...
Sorry, but this is a 'wishful thinking' only.
I was playing arround a lot with the fillament of my 833...
My experience says there are two ways only:
1. Two 5V switching supplys
or
2. A Transformer in the right dimension (11-12V only) whith a huge C/L/C (+R for the rigth Voltage of 10V) without any regulation
But this are MY expereiences only.
Matthias
As hpeter wrote here: http://www.diyaudio.com/forums/tube...e-crisis-my-833c-amp-build-7.html#post3431529
This are 50W, only 😉,...+30W for the choke...
Sorry, but this is a 'wishful thinking' only.
I was playing arround a lot with the fillament of my 833...
My experience says there are two ways only:
1. Two 5V switching supplys
or
2. A Transformer in the right dimension (11-12V only) whith a huge C/L/C (+R for the rigth Voltage of 10V) without any regulation
But this are MY expereiences only.
Matthias
At this point, I'm not worried about it. If there's heat the heat sink/fan can't take care of, the 10mm thick aluminum thermal spreader will also be conducting heat into the chassis top panel (18"x22") and I have lots of room on top to add on another heat sink if needed.
I'm more worried about the 40+Watts the shunt regulator in the 833 cathode is dumping, but again, I can add a lot more heat sink to the side panel where that chip is mounted.