The updated part list for the power supply board (R10 added, example part number for C2 C3 C8 C9 corrected):
| Qty | Value | Device | Package | Parts | Example P/N |
|---|---|---|---|---|---|
| 1 | 2x18V | Power transformer | BLOCK FL 10VA, Talema 15VA | TR1 | See post #174 |
| 1 | 10mH | Common mode choke | LS=7x8 or 10x13mm | L1 | SU9VD |
| 1 | LM317T | LM317T | TO220 | LM317 | LM317T |
| 1 | LM337T | LM337T | TO220 | LM337 | LM337T |
| 2 | DB104 | Diode bridge | DIP4 | BR1, BR2 | DB104 |
| 6 | 1N4148 | Diode | DO35 | D1, D2, D3, D4, D5, D6 | 1N4148-T26A |
| 1 | 47n | X2 safety capacitor | Radial LS=15mm | C1 | 890324025009CS |
| 4 | 150nF | Ceramic or MKS/MKT film capacitor | Radial LS=5mm | C2, C3, C8, C9 | MKS2C031501A00KSSD |
| 2 | 10nF | Ceramic or MKS/MKT film capacitor | Radial LS=5mm | C4, C5 | MKS2G021001A00MSSD |
| 2 | 4700u 25V | Aluminum electrolytic capacitor | Snap-in D=25mm | C6, C7 | B41231B7478M000 |
| 4 | 22u 50V | Aluminum electrolytic capacitor | Radial D=6.3mm | C10, C11, C12, C13 | EEUFC1H220 |
| 2 | 10uF 50V | Aluminum electrolytic capacitor | Radial D=5mm | C14, CLED | EEUFC1H100L |
| 2 | 22 | Resistor 1W | D=2.8mm L=9.0mm | R1, R2 | ROX1SJ22R |
| 2 | TBD | Resistor 0.25W | D=2.5mm L=6.3mm | R3, R4 | See text |
| 2 | 160 | Resistor metal film 0.25W | D=2.5mm L=6.3mm | R5, R7 | MFR-25FBF52-160R |
| 2 | 2k | Resistor metal film 0.25W | D=2.5mm L=6.3mm | R6, R8 | MFR-25FBF52-2K |
| 1 | 1k | Resistor 0.25W | D=2.5mm L=6.3mm | R9 | MFR-25FBF52-1K |
| 1 | 6k8 | Resistor 0.25W | D=2.5mm L=6.3mm | R10 | MFR-25FBF52-6K8 |
| 1 | 1k8 | Resistor 0.25W | D=2.5mm L=6.3mm | RLED | MFR-25FBF52-1K8 |
| 1 | 2-way | Connector | 0.1in contact spacing | LED | 22-23-2021 |
| 1 | 2-way | Connector | 0.312in or 0.4in contact spacing | AC_MAINS | See text |
| 1 | 5-way | Connector | 0.1in contact spacing | OUT2 | 22-23-2051 |
| 2 | 1in tall | Heatsink | Aavid 531x02 | n/a | 531002B02500G |
| 4 | n/a | TO220 thermal pad | n/a | n/a | 43-77-9G |
| 2 | n/a | M3x8 round head screw | n/a | n/a | n/a |
| 2 | n/a | M3 nut | n/a | n/a | n/a |
| 4 | n/a | M3 washer | n/a | n/a | n/a |
| 2 | n/a | M3 spring washer | n/a | n/a | n/a |
- As noted above, the height of the assembled board depends on the height of the transformer (see post #174), two heatsinks, and two large electrolytic capacitors (C6 C7). It makes sense to choose all these parts with approximately the same height, e.g. 1 inch / 25mm. Note that C6 and C7 have max allowable diameter 25mm.
- The common mode choke (L1) can be just about any of the many similar parts with inductance between 2 and 22mH. The PCB accepts chokes with pins in 10x13mm or 7x8mm rectangle.
- The AC input connector can be either a Molex KK396 or KK508 3-way header with middle pin removed, or a three-way 0.2 in pitch terminal block without its middle contact. Of course, soldering bare wire to the PCB is also possible.
- Only one output connector (the 5-way OUT2) is needed for Omicron.
- The value of R1 R2 can be adjusted between 10 and 100 ohm should the voltage drop on the regulators under full load is too small or too big.
- The value of the snubber resistors R3 R4 should be chosen based on the transformer used, e.g. with the help of Mark Johnson's Quasimodo test jig. To give you an example, I found the optimal value was 39 ohm for AMVECO 70064K toroidal transformer and 470 ohm for BLOCK FL10/18 UI type.
- Resistors R5 R6 R7 R8 set the output voltage for the regulators, The value shown in the table set 17V rails. Should you want lower voltage, you need to choose these resistors accordingly. Some popular values would be R5=R7=150 ohm, R6=R8=1.3 kOhm for 12V and R5=R7=200 ohm, R6=R8=2.2 kOhm for 15V,
- Thermal pads are optional as the heatsinks are not connected to anything on the board.
The stock version of Omicron works from +/-17V rails, which allows it to swing its output to about +/-15V and comfortably deliver sufficient power into higher-impedance cans, such as the 600-ohm Beyerdynamic T1. Today, the 600-ohm T1 is more of an exception; the majority of modern can have a lower impedance.
While the stock Omicron is "omnivorous" and works exceptionally well with both high- and low-impedance headphones, for exclusively lower impedance loads, it may make sense to lower the rails and increase the quiescent current of Omicron's output stage. With lower impedance loads, the modified Omicron will stay in Class A at higher signal levels compared to the stock version. I am not sure if there would be an audible difference - try it and report the results!
The option of lowering the supply rails and increasing the quiescent current has been mentioned a couple of times earlier in this thread. Here is how to do it for +/-12V supply.
If you want to use the stock power supply board for lower rails, you'd need to change the dividers that set the regulators' output voltage (R5 R6 and R7 R8, see the schematic in post #300, from 180ohm+2.2kOhm to something like 150ohm+1.3kOhm) and perhaps choose the power transformer with lower output voltage, e.g. 2x15VAC instead of 2x18VAC.
While the stock Omicron is "omnivorous" and works exceptionally well with both high- and low-impedance headphones, for exclusively lower impedance loads, it may make sense to lower the rails and increase the quiescent current of Omicron's output stage. With lower impedance loads, the modified Omicron will stay in Class A at higher signal levels compared to the stock version. I am not sure if there would be an audible difference - try it and report the results!
The option of lowering the supply rails and increasing the quiescent current has been mentioned a couple of times earlier in this thread. Here is how to do it for +/-12V supply.
- Change the value of R11 R12 R31 R32 (see the schematic attached to the posts #295 and #297) to 6.8kOhm - this makes the current through them about the same with 12V rails as it was with 17V and 10kOhm.
- Increase the value of R11 R12 R31 R32 by about 50%. For example, the the SMT version, mount 100 ohm resistors in place of 68 ohm. This increases the quiescent current of the output stage by 50%, keeping the total dissipation with 12V rails the same as it was with higher rails and lower current.
- Short R58 and, for the SMT version only, R43. These resistors drop the voltage in excess of 24V for the relay(s) and are not needed for +/-12V rails.
- Optionally, increase the value of R53 R54 from 100 ohm to 150 ohm to keep the DC protection threshold the same. You can skip this change and live with lower thresholds, it's not a big deal.
If you want to use the stock power supply board for lower rails, you'd need to change the dividers that set the regulators' output voltage (R5 R6 and R7 R8, see the schematic in post #300, from 180ohm+2.2kOhm to something like 150ohm+1.3kOhm) and perhaps choose the power transformer with lower output voltage, e.g. 2x15VAC instead of 2x18VAC.
Last edited:
I need help on a few power supply parts. I'm trying to get these all from mouser to save on extra shipping costs but will gladly order from wherever.
1. common mode choke; I can't find any that meet the mounting parameters.
2. LM317T; 6 types available. Any recommendations on which would be preferred?
511-LM317T; 926-LM317T; 926-LM317T/NOPB; 926-LM317T/LF01; 511-LM317T-DG; 595-LM317KCSE3
3. LM337T; 2 available.
926-LM337T/NOPB; 926-LM337T/LF01
4. c6, c7 cap; 1 available, but it's 25.4mm. Will that just barely fit? If not, is there an acceptable replacement? There is a nichicon that meets the size and capacitance, but 35V instead of 25V; 647-LKG1V472MESAAK
Thank you in advance. I'm still new enough to know not to make assumptions so I appreciate any guidance here.
1. common mode choke; I can't find any that meet the mounting parameters.
2. LM317T; 6 types available. Any recommendations on which would be preferred?
511-LM317T; 926-LM317T; 926-LM317T/NOPB; 926-LM317T/LF01; 511-LM317T-DG; 595-LM317KCSE3
3. LM337T; 2 available.
926-LM337T/NOPB; 926-LM337T/LF01
4. c6, c7 cap; 1 available, but it's 25.4mm. Will that just barely fit? If not, is there an acceptable replacement? There is a nichicon that meets the size and capacitance, but 35V instead of 25V; 647-LKG1V472MESAAK
Thank you in advance. I'm still new enough to know not to make assumptions so I appreciate any guidance here.
^ No one's posted back... I'd suggest that it's time to start opening some datasheets / spec sheets. It's not as bad as it might seem.
1. Be sure you're looking at the mounting specs in the datasheet vs. perhaps the overall dimensions of WxH that some retailers put in their spreadsheet when displaying search results.
2. Filter out all 'duplicates' by looking at the data sheet. Determine if the suffixes and small differences are just packaging, lead free vs. non-lead free, the exact same part with a different number (it happens), or a potential difference in the actual spec of the part. Then, if there are any meaningful differences in the parts or any confusion, you can ask... does it matter if parameter X is A vs. B. Typically if the circuit designer wants you to pick a specific part, they'll be very specific. Not saying that to be fussy. It just might be appreciated if you narrow things down. I won't type in 6 part numbers and scan data sheets, but I can tell that some are functionally identical to others.
3. Same as 2.
4. Looking at the board render. Does it look like an extra 0.4mm will fit? You can always go up in voltage rating.
You'll start to pick up really quickly where to look in data sheets for what all the different numbers / letters within the same part family mean. It's a pain at first, but it's very valuable to learn, IMO.
Lastly, Mouser et. al. change suffixes all the time. If you check the manufacturer's part number, you're always safest, IMO.
Hope that helps and that you find all your parts very quickly. If not, post back. It's really good to not make assumptions, we've all been there.
First data sheet attached for your perusal, since I looked up that part anyway.
Mainly, enjoy the build and have fun!
1. Be sure you're looking at the mounting specs in the datasheet vs. perhaps the overall dimensions of WxH that some retailers put in their spreadsheet when displaying search results.
2. Filter out all 'duplicates' by looking at the data sheet. Determine if the suffixes and small differences are just packaging, lead free vs. non-lead free, the exact same part with a different number (it happens), or a potential difference in the actual spec of the part. Then, if there are any meaningful differences in the parts or any confusion, you can ask... does it matter if parameter X is A vs. B. Typically if the circuit designer wants you to pick a specific part, they'll be very specific. Not saying that to be fussy. It just might be appreciated if you narrow things down. I won't type in 6 part numbers and scan data sheets, but I can tell that some are functionally identical to others.
3. Same as 2.
4. Looking at the board render. Does it look like an extra 0.4mm will fit? You can always go up in voltage rating.
You'll start to pick up really quickly where to look in data sheets for what all the different numbers / letters within the same part family mean. It's a pain at first, but it's very valuable to learn, IMO.
Lastly, Mouser et. al. change suffixes all the time. If you check the manufacturer's part number, you're always safest, IMO.
Hope that helps and that you find all your parts very quickly. If not, post back. It's really good to not make assumptions, we've all been there.
First data sheet attached for your perusal, since I looked up that part anyway.
Mainly, enjoy the build and have fun!
Last edited:
Sorry I am late and thank you @ItsAllInMyHead!
1. There are plenty of suitable common mode chokes, for example: one, two, three, four.
2. Any variant of LM317 in TO-220 will work fine.
3. Any variant of LM337 in TO-220 will work fine.
4. LKG1V472MESAAK is a good choice as is the B41231B7478M000 shown in the table above; BTW both are 35V. There is no practical difference between 25mm and 25.4mm.
Generally, there would be quite a few substitutes for any given part, because the part list for Omicron was designed that way. Looking up a specific part number is the easiest, which I why I published them. Unfortunately, what was available yesterday is sometimes not available today. If the specific part number is unavailable, look for the same parameters (for example, capacitance and rated voltage) and the same size/package/mounting pattern so that the part fits the board. Both Mouser and DigiKey have helpful search engines that allow you to specify everything important and spit out a list of available parts. This works well with resistors, capacitors and most semiconductors, but for whatever reason not so well with chokes or connectors, so you may need to dig a little to find what you need.
1. There are plenty of suitable common mode chokes, for example: one, two, three, four.
2. Any variant of LM317 in TO-220 will work fine.
3. Any variant of LM337 in TO-220 will work fine.
4. LKG1V472MESAAK is a good choice as is the B41231B7478M000 shown in the table above; BTW both are 35V. There is no practical difference between 25mm and 25.4mm.
Generally, there would be quite a few substitutes for any given part, because the part list for Omicron was designed that way. Looking up a specific part number is the easiest, which I why I published them. Unfortunately, what was available yesterday is sometimes not available today. If the specific part number is unavailable, look for the same parameters (for example, capacitance and rated voltage) and the same size/package/mounting pattern so that the part fits the board. Both Mouser and DigiKey have helpful search engines that allow you to specify everything important and spit out a list of available parts. This works well with resistors, capacitors and most semiconductors, but for whatever reason not so well with chokes or connectors, so you may need to dig a little to find what you need.
Last edited:
Another Omicron is born!
This one is a no compromise version, where all signal parts are as far away from the AC parts as possible, in a fairly oversized box.
This one also includes a XFEED switch, and engaging the switch also appears to boost bass.
The case is plain compared to the Fallout 4 edition, but it's Omicron at its best.
Thanks, @alexcp and @Rus2000 for sharing the goods! Lady Fortuna is very pleased.
I'm finishing up the board now and I have a question about bd139/b140. With the isolated packages how can I tell which side to mount to the heatsink? Also it looks like there's a difference in which spot to put them on the written BOM and on the BOM that I imported to mouser. Could someone clarify the correct locations?
The front side of the isolated package is marked with the manufacturer's logo, device type (e.g. "BD139-16") and date code. The back side, which touches the heatsink, is unmarked. If you look at the package from the top or bottom, it is a little wedge-shaped, with the back being a bit broader than the front. Also, if you look at the package from the side, the leads are closer to the front than to the back.
On the board, BD139 are facing the large electrolytic caps and are marked Q1 and Q21. BD140 are facing the opamps and are marked Q2 and Q22.
On the board, BD139 are facing the large electrolytic caps and are marked Q1 and Q21. BD140 are facing the opamps and are marked Q2 and Q22.
Perfect thank you. I did it correctly. I'm just waiting on the magnet wire and the proper connectors for the boards.
If it's not too much trouble would it be possible for you to further explain the correct grounding and wiring scheme. Is the center pin on the input and output of the board ground? Also I'd need to connect the iec inlet to a safety ground, correct? And what would be a good rating for the fuse?
If it's not too much trouble would it be possible for you to further explain the correct grounding and wiring scheme. Is the center pin on the input and output of the board ground? Also I'd need to connect the iec inlet to a safety ground, correct? And what would be a good rating for the fuse?
The center (ground) pins on booth input and output connectors, as well as the two ground pins on the power connector, are all connected together on the board. You don't need to, and shouldn't, connect them off board. Just connect the input connector to the input jacks (with a volume control pot in between if used), output connector - to the headphone jack, and power - to the power supply board.
As with any device using unbalanced (RCA) connection. double insulated construction is preferred to safety grounding the case.
The basic idea of safety grounding is that when insulation fails, the (large) current flows to safety ground and blows the fuse or trips the breaker in the electrical panel. To achive that, you connect every externally accessible part (including the conductive case/chassis) to safety ground. However, to prevent RF ingress, you want to connect the case to the circuit ground, too. The problem is that tying the audio ground to safety ground causes hum problems or more subtle sonic issues. One workaround it to connect the case to safety ground via a ground loop breaker that normally works like open circuit but still allows the fuse-blowing current to flow when needed. A beefy diode bridge is one option:
In this case a fuse on the hot wire is mandatory. Driven at full tilt by a sinewave into a low impedance load, Omicron draws up to about 10VA, and much less on music, so a 100mA or 125mA fuse would be adequate.
BTW Omicron is relatively insensitive to RF ingress (no exposed high impedance points, relatively low input impedance, inputs and outputs protected by RF filters), so you might be able to get away with some other approach, e.g. floating the amp inside a grounded chassis. I have not tested this and am not sure if that's code compliant.
A better approach is to use double insulation (a.k.a. Class II). The basic idea of double insulation is that when one layer of insulation fails, there is a second layer that protects you. A single layer of insulation is one layer of insulating material around a conductor or 3 mm (⅛ in) of air between an uninsulated live part and any externally accessible part. You should be able to guarantee these 3mm even if you pull on the internal wires. Double insulation, then, is two separate layers of insulating material; or one layer of insulating material plus 3mm of air; or 6mm of air. That includes the inside of the power transformer and PCB trace clearance. BLOCK FL transformers are safety class II, and it is not too difficult to make primary wiring double insulated. Then you can use IEC C8 inlet instead of the usual C14, forget about safety grounding and connect the case to the signal ground (at one point only). You probably can omit the fuse, too, as the two primary side resistors (R1 R2) in the power supply will play that role in case of overload. If you like, you can even install fusible resistors into those positions.
As with any device using unbalanced (RCA) connection. double insulated construction is preferred to safety grounding the case.
The basic idea of safety grounding is that when insulation fails, the (large) current flows to safety ground and blows the fuse or trips the breaker in the electrical panel. To achive that, you connect every externally accessible part (including the conductive case/chassis) to safety ground. However, to prevent RF ingress, you want to connect the case to the circuit ground, too. The problem is that tying the audio ground to safety ground causes hum problems or more subtle sonic issues. One workaround it to connect the case to safety ground via a ground loop breaker that normally works like open circuit but still allows the fuse-blowing current to flow when needed. A beefy diode bridge is one option:
In this case a fuse on the hot wire is mandatory. Driven at full tilt by a sinewave into a low impedance load, Omicron draws up to about 10VA, and much less on music, so a 100mA or 125mA fuse would be adequate.
BTW Omicron is relatively insensitive to RF ingress (no exposed high impedance points, relatively low input impedance, inputs and outputs protected by RF filters), so you might be able to get away with some other approach, e.g. floating the amp inside a grounded chassis. I have not tested this and am not sure if that's code compliant.
A better approach is to use double insulation (a.k.a. Class II). The basic idea of double insulation is that when one layer of insulation fails, there is a second layer that protects you. A single layer of insulation is one layer of insulating material around a conductor or 3 mm (⅛ in) of air between an uninsulated live part and any externally accessible part. You should be able to guarantee these 3mm even if you pull on the internal wires. Double insulation, then, is two separate layers of insulating material; or one layer of insulating material plus 3mm of air; or 6mm of air. That includes the inside of the power transformer and PCB trace clearance. BLOCK FL transformers are safety class II, and it is not too difficult to make primary wiring double insulated. Then you can use IEC C8 inlet instead of the usual C14, forget about safety grounding and connect the case to the signal ground (at one point only). You probably can omit the fuse, too, as the two primary side resistors (R1 R2) in the power supply will play that role in case of overload. If you like, you can even install fusible resistors into those positions.
Last edited:
Well I tried to order parts for the smt version, right off the bat the pulse transformer has 0 stock and is end of life, the 8 caps are zero with no substitute. I am at my second home using a tablet so it is hard. I hope this project does not end up in the bin due to parts availability. This has become a problem in the commercial sector as well, I have Schitt headphone amp on order and it too is delayed because they can’t get parts. If I were a conspiracy believer I would suspect the market is being manipulated so only Chinese manufacturers have access to parts.
Sorry you are having troubles with the BOM, parts are awful these days. I did a little part subbing off the bat, here's what I subbed, I have built 2 so far, one with NE5532, one with OPA1642, and the part subbed below. They both sound phenomenal.
C13, C14, C33, C34 - UHV1E102MPD1CC - Lots of stock at mouser
C11, C12, C31, C32 - EEU-FC1E101S - Also lots of stock at mouser, the BOM had 47u but the schematic showed 100u, hence the choice here
C9, C10, C29, C30 - UKW1V470MDD - Plenty at mouser, but the BOM part is also in stock at digikey
L1 - 78602/3MC-R - I did no subbing here, digikey has 400+ in stock
Hope this all helps!
C13, C14, C33, C34 - UHV1E102MPD1CC - Lots of stock at mouser
C11, C12, C31, C32 - EEU-FC1E101S - Also lots of stock at mouser, the BOM had 47u but the schematic showed 100u, hence the choice here
C9, C10, C29, C30 - UKW1V470MDD - Plenty at mouser, but the BOM part is also in stock at digikey
L1 - 78602/3MC-R - I did no subbing here, digikey has 400+ in stock
Hope this all helps!
I like the write up on grounding, in power amplifiers I always used a ground lift scheme but in a low power application like headphone amps it did seem to be overkill. I like to make my cases compact and this will help, if I was more motivated I would learn to 3d print cases and that would add another layer of safety with a non conductive case. I did try to source double insulated wire once from local sources and no one had any and most online vendors required min of 100’ , so I have resorted to keeping the pieces I trim off transformers.
Bill
Bill
Last edited: