AMB Laboratories DIY Audio

[cobretti]

Yes.

[bssanders]

linux-works: a very clever idea and use. Tell me just how you use it to test power amps?

[linux-works]

you can certainly test the power supply on the amp, directly. either with the amp connected or just the psu and this tester.

the polarity does matter and so this active load won't work for ac sources as-is. a friend gave me a hint that you could put a diode bridge between the ac source and this tester and that might work. I'll have to try it to confirm. you do lose .6v per diode when you use the bridge, of course.

I'm just now learning about these boxes that youc an buy called programmable-loads or active-loads. I don't know if they are polarity sensitive or not or how similar they are to this scheme. I'm not going to pay hundreds of dollars to buy one, though!

[reod]

Sorry, I don't have the equipment to do any comprehensive tests.

As far as temperature goes, I've played music on it for several hours continuously and it definitely runs hotter than the AB version, but I can still touch the heat sinks for quite a while before it"feels" hot

[linux-works]

a quick 'build' to test out a new device (new to me). its a fancy high-precision voltage reference that is stable to 1ppm/deg-C, lm399H:

the chips that I got from ebay (some old, some even older):



and a dumb little test circuit, sending in 12v and getting out the pre-set ref voltage (the actual value is not important; what matters is how temperature stable it is). the fluke reading with relative mode on main and absolute mode on 2nd display:



it has 4 leads: 2 for a heater (12v or more) and 2 for the zener. like any zener, you'd add a series R to current limit. I put some caps on the output just to lower any noise (if any). its showing that its stable down to 100uV, which is at the limit of my meter.

I plan to use this where there is an A/D stage or digi-pot or something that I need a stable voltage on. maybe even a programmable power supply...

the little white caps are air covers (really!) that stop air currents from affecting the metal can inside. they are sealed (not air tight) and rattle a bit since they are loose fitting around the bottom wires. neat little things, though!

[amb]

Those LM399s are not that old... they have the most recent National Semiconductor logo.

National's "really old" logo looks like this:


The "old" logo:


The most recent logo:

[cobretti]

Here is my current sink. Works fine on first turn ON. Just have to verify accuracy tomorrow. Pass transistor is mounted on the bottom of the box. I am able to adjust current by 1mA up to....? I don't know yet because the PSU I am using for test (designed by KT88 member) has current limit set to approx. 1.7A, then the overcurrent protection kicks in.

[linux-works]

nice. and cleanly done.

that's a voltmeter/lcd? not a custom cpu and lcd?

[cobretti]

Thanks,
Yes, it is a panel voltmeter , 2V.

[linux-works]

uhm, LED-duino, anyone?



a mostly-headless back-end volume control (and other things) arduino server. it will take serial commands on its uart (or SoftSerial port) and convert them to binary i2c or spi sequences for delta1 relays, delta2 relays, PGA chips and other new solid state volume engines (protocols to be added).

this is a slight fork of the lcduino project, splitting the single image (program) in half and putting the user interface on the lcduino (the current one in AMBs store) and the back-end device code on a board like this.

for extra fun, I included a 10segment bargraph that can be addressed or used separately (or ignored).

board size was chosen to be exactly the same as an lcd display with the same hole pattern. the led bargraph was placed vertically in the center of where the lcd display would have been. the idea is that you could swap out the lcduino in an a10 (for example) and put this headless version in, and have remote remote control ability. the bargraph is very much optional and on embedded builds it would probably be omitted.

one more item: that transistor, diode and resistor cluster is a new protection feature. its meant to stop the wrong firmware from running on the wrong board. since I'm going to be going multi-cpu, its important to make sure that if the user uploaded the GUI firmware to the ENGINE board, nothing bad will happen

project is just starting and this is an early proto.

connectors are: standard ftdi software serial upload (with red sparkfun dongle installed for this photo); a nearly identical row of male headers that could take a 2nd FTDI cable but this is meant for the user to talk over, not for software uploads. it is a 'software serial' or softserial port. it has 1 less pin than the FTDI firmware port since that one needs a reset line for the whole cpu and the user serial port does not need or want that. so, a 5pin for user and 6pin for software upload.

there is the familiar 'amb standard' 6 pin i2c connector: 2 for power, 2 for gnd and 1 for clock and data. that 6pin inline is what we use on the delta1, delta2 and lcduino, so its sort of our standard i2c connector, these days.

new is a 2 pin molex and resistor key. much more detail on this later; but its the firmware protection feature. I might end up calling it 'Firmware Compatibility Check' or the FCC (lol). when the system first boots, it asks the FCC if the chip is on the right board (so to speak) and if things are good, the green led comes on and the code continues to run. if there was a mismatch, the led blinks and code ends there; telling the user to upload the other firmware instead of the one he mistakenly put here


anyway, soon, you'll be able to 'telnet to your volume control'. sort of