My Tuba is built and working. So now I have Qty 2 extra regulators. Mouser p/n 511-LD1084V
They're still out of stock. I want these to go into a someone's build, not a parts stash.
I will send someone the pair for FREE to the first person who sends me a direct message / "Start a conversation" and meets the following criteria:
You're in the USA
You have your Tuba Board(s)
You will have your Tuba filter complete (built, installed/tested) by end of January.
They're still out of stock. I want these to go into a someone's build, not a parts stash.
I will send someone the pair for FREE to the first person who sends me a direct message / "Start a conversation" and meets the following criteria:
You're in the USA
You have your Tuba Board(s)
You will have your Tuba filter complete (built, installed/tested) by end of January.
I am close to wiring my (modded) 24V Tuba filter into a Big MoFo build, and have a question about the speaker thump kill. In the absence of power (during warm-up delay and sudden power termination), does the thump kill establish a parallel (shunt) connection between signal +ve out and ground (i.e., parallel to the speaker load)? I figure it it must, but I wanted to check to be sure before wiring a connection from each speaker +ve out to the board.
MTIA, Derek
MTIA, Derek
So I'm testing the LD1084 part of the circuit. Using Mark Johnson's instructions - trims at max clockwise - I measure a 2.2V across TP1 - TP2, which was a bit of a surprise. Turning counter clockwise increases the V drop, as expected.
According to the datasheet the Vout = 1.25V*(1+(2.7K+1Ktrim)/120R)
trim=1K, Vout=39.8V
trim=0, Vout=29V
Input voltage is 36V so I expected 36V-min V drop (maybe 1.25V) at trim=1k.
All resistances verified so it's a bit of a head scratcher. Any ideas?
According to the datasheet the Vout = 1.25V*(1+(2.7K+1Ktrim)/120R)
trim=1K, Vout=39.8V
trim=0, Vout=29V
Input voltage is 36V so I expected 36V-min V drop (maybe 1.25V) at trim=1k.
All resistances verified so it's a bit of a head scratcher. Any ideas?
Attachments
Your PCB is unfamiliar, it's certainly not the one whose Gerbers I attached to post #1 of this thread (image below). First talk to the people who created that board; has anyone else built one and gotten it to work?
Posts #45 and #61 of this thread, exhibit successful builds of the board shown below.
@alazira
Posts #45 and #61 of this thread, exhibit successful builds of the board shown below.
@alazira
Attachments
Last edited:
If you run Tuba all by itself, with no external load, nothing bad happens when you turn the trimmer all the way clockwise or all the way counterclockwise. Why not try that and see whether the measurements are reasonable. (Voltage at point X with respect to Ground). Also check your old Mouser and DigiKey orders to find out what resistor values you actually bought. 2.7K? 2.75K? 3.3K? 3.32K? 120R? 110R? etc.
_
_
Attachments
Update: Success!
With no load connected, 36V input
To achieve Vdrop of 2V I had to increase the 2.7K+trim1K to 3.3K+trim1K. I measured 4.1K for 2V drop.
However, once connected to the amp the Vdrop went to 1.1V, which would appear to be the minimum drop. complete counter clockwise (trim=0) did not move this.
Changed the 3.3K back to 2.7K and reset trim to 1K. Vdrop now 1.7V and tuneable to 2V. There is some drift lower as the part heats up. I measured a max of 53deg C on the heat sinks.
Conclusion: The 1084V part calculations worked with load but not unloaded (at least in my case). Also, I see from the datasheet that the part has a rated max of 30V so here we are running it 20% over. Not sure how this will play out for long term durability. Hopefully they were conservative on the rated max.
With no load connected, 36V input
To achieve Vdrop of 2V I had to increase the 2.7K+trim1K to 3.3K+trim1K. I measured 4.1K for 2V drop.
However, once connected to the amp the Vdrop went to 1.1V, which would appear to be the minimum drop. complete counter clockwise (trim=0) did not move this.
Changed the 3.3K back to 2.7K and reset trim to 1K. Vdrop now 1.7V and tuneable to 2V. There is some drift lower as the part heats up. I measured a max of 53deg C on the heat sinks.
Conclusion: The 1084V part calculations worked with load but not unloaded (at least in my case). Also, I see from the datasheet that the part has a rated max of 30V so here we are running it 20% over. Not sure how this will play out for long term durability. Hopefully they were conservative on the rated max.
You forgot that the LD1084 (and the LM317) voltage regulator IC treats its output pin as the circuit reference. All voltages are with respect to the output pin -- the IC doesn't know and doesn't care where "ground" is.
So the datasheets give maximum permitted input-to-output voltage. Not input-to-ground voltage, not output-to-ground voltage. Input-to-output voltage. Thanks to the Tuba's zener diode, and its choice of trimming range, input-to-output voltage always remains well within the datasheet specs. T.I.'s LM317 datasheet explains this more thoroughly than S.T. Microelectronics' LD1084 datasheet.
This is what allows vacuum tube folks to use LM317's as 160 volt regulators. They carefully control input-to-output voltage, and let output-to-ground go sky high. Check out the famous Linear Brief attached here.
_
So the datasheets give maximum permitted input-to-output voltage. Not input-to-ground voltage, not output-to-ground voltage. Input-to-output voltage. Thanks to the Tuba's zener diode, and its choice of trimming range, input-to-output voltage always remains well within the datasheet specs. T.I.'s LM317 datasheet explains this more thoroughly than S.T. Microelectronics' LD1084 datasheet.
This is what allows vacuum tube folks to use LM317's as 160 volt regulators. They carefully control input-to-output voltage, and let output-to-ground go sky high. Check out the famous Linear Brief attached here.
_
I have added (2) 10,000uF 4-pole DNM caps on the output. The meanwell SMPS did not seem bothered.
I have made a provision for a Wima DC Link 40uF cap at C3 (in Mark's schematic). Would substituting this for the 33uF electolytic alter the filter freq?
I have made a provision for a Wima DC Link 40uF cap at C3 (in Mark's schematic). Would substituting this for the 33uF electolytic alter the filter freq?
Attachments
Tuba C3 and C4 help prevent the LD1084 chips from oscillating; the datasheet says "a minimum capacitor of 10uF is needed for stability". As long as you don't increase their lead length or lead inductance or distance from the ICs, increasing the capacitance beyond 33uF would be fine.
_
_
Attachments
Remember that the Tuba PCB contains two voltage regulators: one for the Left amplifier channel, and another for the Right amplifier channel. This is shown on the schematic attached to post #1.
A stereo amplifier, such as the VFET amps sold by lottery in the diyAudio Store, only needs one Tuba board.
A stereo amplifier, such as the VFET amps sold by lottery in the diyAudio Store, only needs one Tuba board.
Thanks Mark, I am aware of that. I will be building two VFET amps, one P and one N. I already have the VFETs and the two kits from the DiyAudio Store. I also ordered two VFET chassis.
My apologies up front - this is probably an extremely dumb question...
I think I understand how this circuit works and it's very cool. I've ordered one for my VFET OS2 kit. But my question is: Why doesn't shorting the output to ground damage the output section? Other amps I've worked on are easily damaged if the speaker outputs are shorted together.
I think I understand how this circuit works and it's very cool. I've ordered one for my VFET OS2 kit. But my question is: Why doesn't shorting the output to ground damage the output section? Other amps I've worked on are easily damaged if the speaker outputs are shorted together.