LOL. Personally, I don't take SPICE results literally, but I do think it's great for developing intuition about how a circuit works. I found this article just now with a quote from Bob Pease, which is pretty much how I think about SPICE:
About the Tian probe, in case my comments above were too cryptic. To simulate loop stability, you need to check the response from the input stage to the output stage, and then all the way back through the feedback network to the input again. This requires breaking the loop to measure it. But the amp typically doesn't work with the feedback loop disconnected. So we insert a "probe" inside the loop that keeps the feedback intact while giving us an access point for measuring the voltages and currents. Then, with some math, we can derive the loop transfer function.
I'm not gonna lie. Sound quality may be the final measure of success. But I'm an engineer and I enjoy the technical theory. I'm incredibly amused that I built something as complicated as the A2, and that it worked first time, and sounds really good (to my tin ears).
About the Tian probe, in case my comments above were too cryptic. To simulate loop stability, you need to check the response from the input stage to the output stage, and then all the way back through the feedback network to the input again. This requires breaking the loop to measure it. But the amp typically doesn't work with the feedback loop disconnected. So we insert a "probe" inside the loop that keeps the feedback intact while giving us an access point for measuring the voltages and currents. Then, with some math, we can derive the loop transfer function.
I'm not gonna lie. Sound quality may be the final measure of success. But I'm an engineer and I enjoy the technical theory. I'm incredibly amused that I built something as complicated as the A2, and that it worked first time, and sounds really good (to my tin ears).
Well, the 500 Ohm trim pots are almost gone. Arrow had three so I scored two for four dollars and six dollars shipping, ouch.
The SOIC-8 LM555 is out of stock at Mouser. Digi-Key has some, but fortunately I found extras in my parts pile. The Molex 2-pin connector housings are also out of stock, but again I found some in the closet. It looks like I can get everything I need to build up the board. I should submit my order now, before more things go out of stock.
The SOIC-8 LM555 is out of stock at Mouser. Digi-Key has some, but fortunately I found extras in my parts pile. The Molex 2-pin connector housings are also out of stock, but again I found some in the closet. It looks like I can get everything I need to build up the board. I should submit my order now, before more things go out of stock.
FWIW, I got a box full of parts from Mouser today, and this morning I sent Toroidy $190 for a custom power transformer. I asked for 100VA, so even Mark should be satisfied. Boards are in New York City right now.
Economical packaging.
Edit: It's good they're thinking about protecting the sensitive wiper-track junctions from ESD.
My boards are out for delivery today... Exciting.
Speaking of ESD, I blew a slew of IRF MOSFETs last winter just handling them. They seemed to work in circuit, but failed to identify on my Peak Atlas tester, and showed excessive gate leakage with the DMM. This was in spite of taking reasonable care, with a static dissipating pad on the bench.
I need to get a humidifier for my workroom. My current protocol is to take the parts carefully out of the bags and immediately stick the leads through aluminum foil, which I leave in place until they are soldered to the board.
Speaking of ESD, I blew a slew of IRF MOSFETs last winter just handling them. They seemed to work in circuit, but failed to identify on my Peak Atlas tester, and showed excessive gate leakage with the DMM. This was in spite of taking reasonable care, with a static dissipating pad on the bench.
I need to get a humidifier for my workroom. My current protocol is to take the parts carefully out of the bags and immediately stick the leads through aluminum foil, which I leave in place until they are soldered to the board.
Henry if you have not already I suggest you invest in an ESD bench mat....saved me a lot of grief. ESD can erode connections inside mos and cmos devices and can cause fail after a period of time. Laser assemblies for cd and dvd players are particularly vulnerable.
Please post pics of boards when they arrive.
Jam
Please post pics of boards when they arrive.
Jam
I'm like, "This can't possibly work." But it probably will. Hopefully.
I have the parts, but probably won't start building the boards until the weekend. Once I know the board is working, I will keep one more for myself and offer the remaining three to interested parties as a token of my gratitude. One is already spoken for...
I have an ESD mat but I wasn't using the wrist strap last winter. I'm not sure what went wrong, but I plan to be super careful.
Oh yes. I'm looking forward to how this iteration goes together; still applauding from the quiet margins of the thread!
ATB
ATB
I made some progress tonight. The relay timer works. I need to match transistors, then will finish up the first channel tomorrow and try to test it.
Soldering SMD components isn't that hard. Get the following:
1. Fine-tipped tweezers
2. Syringe of paste flux with needle tip
3. Thin silver-bearing solder
4. Fine-tipped soldering iron
5. Solder wick
6. Flux remover
7. Toothbrush, cotton swabs, paper towels
8. Magnifying goggles
To install the parts, apply flux to the pads. Put a small amount of solder on the iron tip. Place the component on the pads and align in place with the tweezers. Touch the solder-wetted iron tip to the junction of the pad and component and flow solder onto the joint. Press the component down with the tweezers and reapply heat so it sits flush. If the alignment isn't right, adjust the part. Then solder the other pad(s). If you get too much solder on the pad, remove some with solder wick and try again.
For SOIC parts, you can just run the wet iron down the row of pins and the solder will flow on nicely. With a little practice, you get a feel for how to control the amount of solder you apply. Use plenty of flux. You can remove bridges with the solder wick, but I rarely have to do that. When satisfied, use flux remover and your choice of tool to clean off the flux. I have been using denatured alcohol stove fuel from the hardware store for flux remover. It works perfectly and is much cheaper than the purpose-bottled stuff.
Another trick for cleaning boards: I fold up a half paper towel until it's about 1" x 2". I soak the folded end with alcohol and scrub the board over a waste barrel. The trimmed leads shred the towel, taking away the flux residue with the shredded bits and exposing fresh alcohol. Finally, remove the paper debris with the toothbrush.
I bought a humidifier this morning. It's 15F outside, but I've got 55% relative humidity in the room, which should help with ESD.
I've installed all the small-signal transistors just now. I need to drill a sheet of scrap aluminum so I can mount the boards and position the output transistors properly for soldering. It's damned cold in the garage where the drill press is.
1. Fine-tipped tweezers
2. Syringe of paste flux with needle tip
3. Thin silver-bearing solder
4. Fine-tipped soldering iron
5. Solder wick
6. Flux remover
7. Toothbrush, cotton swabs, paper towels
8. Magnifying goggles
To install the parts, apply flux to the pads. Put a small amount of solder on the iron tip. Place the component on the pads and align in place with the tweezers. Touch the solder-wetted iron tip to the junction of the pad and component and flow solder onto the joint. Press the component down with the tweezers and reapply heat so it sits flush. If the alignment isn't right, adjust the part. Then solder the other pad(s). If you get too much solder on the pad, remove some with solder wick and try again.
For SOIC parts, you can just run the wet iron down the row of pins and the solder will flow on nicely. With a little practice, you get a feel for how to control the amount of solder you apply. Use plenty of flux. You can remove bridges with the solder wick, but I rarely have to do that. When satisfied, use flux remover and your choice of tool to clean off the flux. I have been using denatured alcohol stove fuel from the hardware store for flux remover. It works perfectly and is much cheaper than the purpose-bottled stuff.
Another trick for cleaning boards: I fold up a half paper towel until it's about 1" x 2". I soak the folded end with alcohol and scrub the board over a waste barrel. The trimmed leads shred the towel, taking away the flux residue with the shredded bits and exposing fresh alcohol. Finally, remove the paper debris with the toothbrush.
I bought a humidifier this morning. It's 15F outside, but I've got 55% relative humidity in the room, which should help with ESD.
I've installed all the small-signal transistors just now. I need to drill a sheet of scrap aluminum so I can mount the boards and position the output transistors properly for soldering. It's damned cold in the garage where the drill press is.
I decided to test the front end before adding the output transistors. I got roughly 6mA in the input stage and 12mA in the VAS stage. I was able to trim the offset for equal OPS gate voltages, and the bias range looks good. This is promising.
Ok, well, it seems to be working. Gain is about 2.6, bandwidth around 1.3 MHz, driven from a low-impedance source.
I biased it to 100mA. During warmup, the bias current drops right off the bat. It's very sensitive to the temperature of the bias spreader transistor; just putting a finger on the transistor increases the bias immediately by tens of milliamps. It seems to have settled down now. I imagine it will be better once it's in an enclosure, but bias will probably vary with ambient temperature (it's cold in this room right now). I can understand why the instructions say to let the amp warm up for an hour and to leave it on all the time. If I were to improve this circuit, I'd consider starting by designing a bias circuit with thermal compensation.
The offset servo seems to work. With the servo disabled, it's possible to trim the offset to within a millivolt or two and it doesn't wander much. The servo is very slow and its action seems quite mild. Offset has stabilized at around 1mV or so.
I need to take a break and then I will start building up the second channel, maybe tonight if I have the energy.