Hey Folks;
It's my turn to give back some info to the group. I know that a few of you have some of the Behringer gear lurking in their closets. The gist of this email - MOST of the Behringer design is VERY solid and VERY WELL DONE! They use a cheap "core" in their products - bad op amps and poorly bypassed supplies. Take care of these things and you have a KILLER product on your hands. Here's how...
For the past several weeks I've been digging into the Behringer T1953 Tube Ultragain Vintager Series unit. What a fun task that was. I learned a LOT of things from this unit that I can share on this forum. Heck Behringer steals from everyone else, why can't I give away a few of their secrets.
THE MODS I DID
There are things that make sense with these units, and things that don't. So, I did the things that made sense, for the time I had to work on it. There are other changes that could be made to this unit, however, and get some incremental gains. My overall goal was to significantly improve performance.
Power Supply RF Bypassing
The power supply is the first place I spent some time. Dudes, this thing is UNDER DESIGNED to say the least. They have a few bypass / smoothing caps in the power supply but they're small, under-valued things. The only way out of this is to actually re-work the entire supply (not on my list of things to do, today).
A second thing that Behringer did, on this thing, was to put most of their "filtering" caps ON THE OTHER PCB'S in the unit!!! Go figure. So, on all the op amps you'll see that on the V+ and V- pins there's a 35V, 10uF cap to ground for "filtering". While this is sort of a good idea overall, they forgot the RF bypass caps on each of the op amps. As a result there was a significant amount of RF noise - "oscillating" the power supplies and causing real noise (audible) in the outputs.
I took all of the power supply bypass caps, on each of the boards and shunted them with a high-quality, 50V, C0G/NP0, military grade 1uF SMD capacitor. This cleaned up the noise considerably - prior to any other mods to the unit.
There is still a small amount of 60Hz hum on the power supply lines. However, the operational amps that I put in the thing have a 130dB PSRR (Power Supply Rejection Ratio) so this wasn't much of a concern to me. The op amps don't reject the high frequency ranges, as much, which is why the bypass caps.
NJM4580 OPERATION AMPLIFIER ISSUES
Behringer uses very low-quality NJM4580 Operational Amplifiers in the Behringer products. I don't care what the product lit says, the 4580 I IS NOT a high-quality, hi-precision operational amplifier. This thing has HORRID settling times (when square waves or transients are sent through the amp they should have flat tops - with a high settling time the top actually has a sine wave in it). This translates into excess noise.
The op amps have a HUGE THD + Noise specification. The higher the frequency the more noise they actually add to the signal. This causes a "harsh" top end and "blurring" of the signal.
The op-amps have a low slew rate - the ability to track the input signal. They don't track well, at all.
The op-amps have a low CMRR (Common Mode Rejection Ratio) - the ability to reject (cancel) a signal that's the same on both inputs.
Noise Figure - the NJM4580's are very high noise, to say the least. This figure refers to how much self-noise the op amps add to the already existing noise from your microphones (active condensers for example), resistors (yes they generate noise), etc.
Bandwidth limitations - the NJM4580's have some pretty severe bandwidth limitations, as well. Interestingly enough, the square wave response of the preamp was very interesting. When a square wave is sent through an audio device the wave should be, well, square, if all the devices in the audio path are configured flat.
On the Behringer, T1953 the square wave was not flat on top it had a rise (pretty good one, at that) from the leading edge (high frequency response) to the trailing edge (low frequency response). When the leading edge of the square wave is lower than the trailing edge this equals poor high frequency response. When the trailing edge is lower than the leading edge it signifies poor low frequency response.
The next step in the upgrade process was to remove all the NJM4580's off the board.
BOARD LAYOUT ISSUES
There are several interesting board layout issues that can never be solved unless one just redesigns the entire board. First, and foremost is power/ground distribution. Long cable runs from the power supply, small power traces and ground traces on the PCB's and most of the power supply filtering caps located a LONG way away from the power supply.
This creates an environment that's very conducive to picking up a great deal of RF/EMI and forwarding that interference directly to your active modules (operational amplifiers).
A second point to note about the power supply filtering caps on the individual PCB's is that they were located far away from the operational amplifiers they were meant to filter. Realistically, they should be pretty close to the operational amplifier - as close as possible.
The audio signal path presented some other interesting layout issues. First, and foremost, the audio signal path, like power supply and ground paths should be as short as possible. Many times, the Behringer folks ran that audio path all over the PCB, around components, and traces, etc. This has the potential to pick up EMI and place it right on your inputs to the active devices in the signal path.
The last layout issue of interest is the fact that Behringer uses dual and quad operational amplifiers across the entire signal path. Interestingly enough, there is a good deal of crosstalk between the individual "channels" in an operational amplifier. They would have been better served to have single op-amp packages. It's cheaper to do what they did, though. This doesn't add a lot of impacts to the overall performance as they, at least, are dedicating a dual op-amp to a single channel and using both stages to amplify the same signal. If they had channel 1 on one side and channel 2 on the other it would have been much worse.
The last "gripe" about the design of this unit is the fact that ALL signals are carried en-masse around this thing in unshielded, separated wire connectors. In reality, there is a bit of CMRR (Common Mode Rejection Ratio) that goes on when the wire pairs are "woven" or "braided" together. Behringer could have done this but they chose not to.
Rather than braid the pairs, myself (which requires removing individual wires from connectors) I put a Belden EMI shield over each of the wire packs. This is a hi-density, RF-shielding (all metal) which slips over the wires. I tied one end of this to the system ground to fully shield all the cable packs running between the power supply and the boards, the input PCB and the Pre Amp PCB and the Preamp PCB and the Tube UTC PCB. This had a small, but significant, impact on the integrity of low-level signals (especially) going from the input jack to the preamp PCB.
HOW BAD IS IT REALLY?
After having worked out some pretty extensive trace/signal analysis the overall impact of some of this wasn't as great as it could have been and it's not as good as it could be, either. In the end, the net effect wasn't bad, overall. There were other issues that contribute to the degradation of this unit is a far more significant manner. The worst offenders are the un-bypassed (RF) filtering capacitors on the power supplies of each of the operational amplifiers and the tubes, the operational amplifiers themselves and the un-bypassed electrolytics in the signal path.
OPERATIONAL AMPLIFIER REMOVAL / REPLACEMENT
This was the most fun portion of the entire exercise. Behringer, rather than perform the tradition "oven" method of soldering the chips to the board decided to put Loctite, or similar glue and glue the operational amplifiers to the board. The WORST case of this was on the input/output PCB where the glue was so thick it oozed out from beneath the op amps.
This glue made it nearly impossible to lift the op amps off the board - they're Surface Mount SOIC's. In the case of one of the op amps we actually had to destroy a pad during the lift-off procedure. Thankfully, it was pretty easy to repair.
On the "pre-amp PCB" (which is mounted on the front of the unit) the removal of the operational amplifiers was much easier. They, too, were glued but not with as much glue making their removal much easier and with no destruction of the pads.
I replaced the NJM4580's with Burr-Brown OPA2228's which are an incredibly low-noise, high-precision operational amplifier. These things are 100 times, in most areas, better than the NJM4580's. All-in-all there are 10 of the 4580's on the two PCB's.
The PCB's are split up very nicely, from a visual perspective, not well from a design perspective. However, as noted above there is excessive length in the signal, and ground paths allowing a bit more EMI/noise to enter the system than what one would like.
The nice thing about the Behringer design is the fact that all the operational amplifiers are oriented the same way making the process MUCH easier. If you got one right, you got them all right.
I enjoyed that immensely.
OPERATIONAL AMPLIFIER REPLACEMENT RESULTS
This was the LARGEST single contributor to the overall performance of the T1953. The amount of noise riding on the output signal was reduced by a factor great than 10... yes, you read that right, the amount of noise riding on the output signal was reduced by a factor great than 10. It cleaned up so nicely that the output wave actually looked like a square wave and not a triangle wave with hair on it, randomly pulsing. It was absolutely amazing how much of a difference those operational amplifiers made.
The output frequency response INVERTED. Previously, the T1953 was high-frequency impaired in that the square wave slanted toward the right. Now, it slanted left (low-frequency response). THIS IS GOOD!!! THIS IS GOOD.
I know, from reviewing Behringer's data sheet that the T1953's low-end 3dB point was 30Hz. The square wave actually showed that, correctly, now. This can be corrected by replacing the caps in the signal path - more on that later. This was the LARGEST SINGLE improvement that could be made to this box. The end result was stunning to say the least.
OTHER SIGNAL PATH UPGRADES
The next step, then was to figure out the last of the signal path upgrades that could be done WITHOUT costing an arm and a leg. I could have replaced all the Electrolytic caps, in the signal path with Black Gate caps. However, they would have cost me as much as the unit did, originally - silly idea, IMO. The goal is the most bang, for the buck - in other words, allowing significant performance improvements for each $$$ spent on the unit. To that end there are a few things that can still be done for the unit and improve performance. First thing you'll notice about this thing is that there are a LOT of electrolytic caps on the preamp PCB. There are 10 operational amplifiers, taking 20 of them for filtering, but there's something like 30 on the board.
First, how to determine the signal path electrolytics from the filter (power supply group). Pretty simple, really. The operational amps are powered with 15VDC. Typically, the rule of thumb is to use a cap with twice the voltage rating of the supply voltage. There are a bunch of 35V caps on the board - these are filter caps for the operational amplifiers.
There are a bunch of 10V caps on the board, too. These are SIGNAL PATH caps. Ahhh... that's how they do that. There are a couple of 60V caps, too. Those are for filtering the 48VDC phantom power.
At this point in time, I didn't have the time to spend to get me some non-polarized electrolytics for the signal path. Truth be told, that would provide some performance improvement, but not having another order ready for DigiKey, right now, made that "upgrade" a bit more $$$ than I wanted to spend - so I didn't do it.
The "problem" with the electrolytic capacitor is that it's really only a capacitor until about 5-6kHz, or so. Then it turns into an inductor and causes all sorts of problems. To rectify this, I got me a bunch more of the 1uF military grade surface mount caps and shunted all of the electrolytics in the signal path. This, again, significantly improved my high frequency performance.
WHAT IS THE UTC CIRCUIT?
I have no idea, really. Actually, interestingly enough, this circuit is a parallel tube (sort of like a parallel effects loop) that the signal is fed into. The signal is grossly distorted, the high-frequency harmonics filtered out and mixed into the signal path. Warmth? Not really, IMO, it works more like an exciter than something that adds warmth. It really adds a bit of "presence" to the signal that wasn't there before.
Interestingly enough, with the NJM4580's in the circuit the effect of the "warmth" wasn't hardly noticeable, even when I cranked the knob all the way to the right and the meter was pegged with all the warmth I was supposed to be adding. Now, however, the effect of the UTC circuit is VERY apparent. It adds a nice touch to the overall signal - especially voice, I've noticed.
THE FINAL RESULTS
That done, I took the unit out of the lab and took it home where I could actually test this thing in a studio environment with some relatively good mic's (RODE NTK, K2 and SHURE KSM32 and KSM44). I A/B'd the unit against my Groove Tubes VIPRE...
I have to say that, overall, the Groove Tubes and the T1953 are pretty hard to differentiate alone - you can tell a bit but not much. When they're in the mix, they both sound the same. The overall clarity, presence, dynamic range and depth of the T1953 had improved so much that it was nearly on par with a $3500 / channel mic pre. There's slightly more noise than the Groove Tubes - but hey, I spent a total of something like $50 in parts on this thing and $139. This is, now, a $200 product that comes close to sounding like a product greater than 10x it's price range.
WHAT UPGRADES MAKE SENSE ON THE BEHRINGER UNITS?
Due to the fact that Behringer uses surface mount components in everything there is some risk every time one performs an upgrade on these boxes.
I started out with high hopes of ordering every surface mount resistor and cap in some highly accurate, highly stable configuration to really improve the thing. I was going to order black gate caps, metalized film polypropelynes, metal film .5% resistors and the whole nine-yards.
The plain fact of the matter is that would have been very dangerous. Each of the surface mount devices has this leetle beety pad beneath it. That pad is the devices connection to every other device in its path. Those pads enjoy removing themselves politely from the PCB when heat is applied. Once they're removed you have to bridge between the component and the pad - making the task significantly more complex. Given the fact that Behringer actually glues every device to the PCB makes the chance of actually removing a pad significantly higher. In fact, on the T1953, I did remove a pad from one of the operational amplifiers. The glue was on the pad and when I removed the operational amplifier the pad came with it. The incremental value that you get for replacing components like this does not, IMO, offset the risk of replacement. The incremental performance upgrade doesn't offset the risk, either.
Upgrades of the signal path capacitors (electrolytics) make sense. Installing a same size, higher voltage (better performance) non-polarized cap makes HUGE sense. Performance improvement is there and it's all good.
Bypassing ALL electrolytics with smaller caps to improve high-frequency response - also a good thing. The surface mount caps really make this process nice. It improves high frequency response in the signal path AND improves RF filtering on the power supply.
Adding higher quality operational amplifiers to the signal path. This is the single greatest improvement in the Behringer product line that you can make. HUGE improvement gains can be had here. Adding FET input operational amplifiers, or other instrument grade operational amplifiers makes a HUGE difference in performance.
Adding RF shielding all the way around the chassis. Especially the cable packs. This makes a huge difference in the performance of low-level signals. With less EMI interfering w/ them, you'll be able to better hear the small transients and differences.
I have not pulled apart my DSP9024 ULTRADYNE, DSP8024 ULTRACURVE, ULTRAFEX PRO or the Virtualizer Pro. I'm betting that the performance of the digital gear can be upgraded, as well. Either with higher performance op amps feeding the A/D converters (that will help), higher performance op-amps feeding the outputs (that will help) better signal path components, better filtering on the power supply (stops A/D and D/A conversion glitches) and possibly even better A/D converters and D/A converters. However, that said, I haven't touched, or ever opened, one of their digital boxes. I will, eventually. I do know, however, that the ULTRAFEX PRO (all analog) will suffer the same performance improvements as what the mic pre's do when they're upgraded.
It's my turn to give back some info to the group. I know that a few of you have some of the Behringer gear lurking in their closets. The gist of this email - MOST of the Behringer design is VERY solid and VERY WELL DONE! They use a cheap "core" in their products - bad op amps and poorly bypassed supplies. Take care of these things and you have a KILLER product on your hands. Here's how...
For the past several weeks I've been digging into the Behringer T1953 Tube Ultragain Vintager Series unit. What a fun task that was. I learned a LOT of things from this unit that I can share on this forum. Heck Behringer steals from everyone else, why can't I give away a few of their secrets.
THE MODS I DID
There are things that make sense with these units, and things that don't. So, I did the things that made sense, for the time I had to work on it. There are other changes that could be made to this unit, however, and get some incremental gains. My overall goal was to significantly improve performance.
Power Supply RF Bypassing
The power supply is the first place I spent some time. Dudes, this thing is UNDER DESIGNED to say the least. They have a few bypass / smoothing caps in the power supply but they're small, under-valued things. The only way out of this is to actually re-work the entire supply (not on my list of things to do, today).
A second thing that Behringer did, on this thing, was to put most of their "filtering" caps ON THE OTHER PCB'S in the unit!!! Go figure. So, on all the op amps you'll see that on the V+ and V- pins there's a 35V, 10uF cap to ground for "filtering". While this is sort of a good idea overall, they forgot the RF bypass caps on each of the op amps. As a result there was a significant amount of RF noise - "oscillating" the power supplies and causing real noise (audible) in the outputs.
I took all of the power supply bypass caps, on each of the boards and shunted them with a high-quality, 50V, C0G/NP0, military grade 1uF SMD capacitor. This cleaned up the noise considerably - prior to any other mods to the unit.
There is still a small amount of 60Hz hum on the power supply lines. However, the operational amps that I put in the thing have a 130dB PSRR (Power Supply Rejection Ratio) so this wasn't much of a concern to me. The op amps don't reject the high frequency ranges, as much, which is why the bypass caps.
NJM4580 OPERATION AMPLIFIER ISSUES
Behringer uses very low-quality NJM4580 Operational Amplifiers in the Behringer products. I don't care what the product lit says, the 4580 I IS NOT a high-quality, hi-precision operational amplifier. This thing has HORRID settling times (when square waves or transients are sent through the amp they should have flat tops - with a high settling time the top actually has a sine wave in it). This translates into excess noise.
The op amps have a HUGE THD + Noise specification. The higher the frequency the more noise they actually add to the signal. This causes a "harsh" top end and "blurring" of the signal.
The op-amps have a low slew rate - the ability to track the input signal. They don't track well, at all.
The op-amps have a low CMRR (Common Mode Rejection Ratio) - the ability to reject (cancel) a signal that's the same on both inputs.
Noise Figure - the NJM4580's are very high noise, to say the least. This figure refers to how much self-noise the op amps add to the already existing noise from your microphones (active condensers for example), resistors (yes they generate noise), etc.
Bandwidth limitations - the NJM4580's have some pretty severe bandwidth limitations, as well. Interestingly enough, the square wave response of the preamp was very interesting. When a square wave is sent through an audio device the wave should be, well, square, if all the devices in the audio path are configured flat.
On the Behringer, T1953 the square wave was not flat on top it had a rise (pretty good one, at that) from the leading edge (high frequency response) to the trailing edge (low frequency response). When the leading edge of the square wave is lower than the trailing edge this equals poor high frequency response. When the trailing edge is lower than the leading edge it signifies poor low frequency response.
The next step in the upgrade process was to remove all the NJM4580's off the board.
BOARD LAYOUT ISSUES
There are several interesting board layout issues that can never be solved unless one just redesigns the entire board. First, and foremost is power/ground distribution. Long cable runs from the power supply, small power traces and ground traces on the PCB's and most of the power supply filtering caps located a LONG way away from the power supply.
This creates an environment that's very conducive to picking up a great deal of RF/EMI and forwarding that interference directly to your active modules (operational amplifiers).
A second point to note about the power supply filtering caps on the individual PCB's is that they were located far away from the operational amplifiers they were meant to filter. Realistically, they should be pretty close to the operational amplifier - as close as possible.
The audio signal path presented some other interesting layout issues. First, and foremost, the audio signal path, like power supply and ground paths should be as short as possible. Many times, the Behringer folks ran that audio path all over the PCB, around components, and traces, etc. This has the potential to pick up EMI and place it right on your inputs to the active devices in the signal path.
The last layout issue of interest is the fact that Behringer uses dual and quad operational amplifiers across the entire signal path. Interestingly enough, there is a good deal of crosstalk between the individual "channels" in an operational amplifier. They would have been better served to have single op-amp packages. It's cheaper to do what they did, though. This doesn't add a lot of impacts to the overall performance as they, at least, are dedicating a dual op-amp to a single channel and using both stages to amplify the same signal. If they had channel 1 on one side and channel 2 on the other it would have been much worse.
The last "gripe" about the design of this unit is the fact that ALL signals are carried en-masse around this thing in unshielded, separated wire connectors. In reality, there is a bit of CMRR (Common Mode Rejection Ratio) that goes on when the wire pairs are "woven" or "braided" together. Behringer could have done this but they chose not to.
Rather than braid the pairs, myself (which requires removing individual wires from connectors) I put a Belden EMI shield over each of the wire packs. This is a hi-density, RF-shielding (all metal) which slips over the wires. I tied one end of this to the system ground to fully shield all the cable packs running between the power supply and the boards, the input PCB and the Pre Amp PCB and the Preamp PCB and the Tube UTC PCB. This had a small, but significant, impact on the integrity of low-level signals (especially) going from the input jack to the preamp PCB.
HOW BAD IS IT REALLY?
After having worked out some pretty extensive trace/signal analysis the overall impact of some of this wasn't as great as it could have been and it's not as good as it could be, either. In the end, the net effect wasn't bad, overall. There were other issues that contribute to the degradation of this unit is a far more significant manner. The worst offenders are the un-bypassed (RF) filtering capacitors on the power supplies of each of the operational amplifiers and the tubes, the operational amplifiers themselves and the un-bypassed electrolytics in the signal path.
OPERATIONAL AMPLIFIER REMOVAL / REPLACEMENT
This was the most fun portion of the entire exercise. Behringer, rather than perform the tradition "oven" method of soldering the chips to the board decided to put Loctite, or similar glue and glue the operational amplifiers to the board. The WORST case of this was on the input/output PCB where the glue was so thick it oozed out from beneath the op amps.
This glue made it nearly impossible to lift the op amps off the board - they're Surface Mount SOIC's. In the case of one of the op amps we actually had to destroy a pad during the lift-off procedure. Thankfully, it was pretty easy to repair.
On the "pre-amp PCB" (which is mounted on the front of the unit) the removal of the operational amplifiers was much easier. They, too, were glued but not with as much glue making their removal much easier and with no destruction of the pads.
I replaced the NJM4580's with Burr-Brown OPA2228's which are an incredibly low-noise, high-precision operational amplifier. These things are 100 times, in most areas, better than the NJM4580's. All-in-all there are 10 of the 4580's on the two PCB's.
The PCB's are split up very nicely, from a visual perspective, not well from a design perspective. However, as noted above there is excessive length in the signal, and ground paths allowing a bit more EMI/noise to enter the system than what one would like.
The nice thing about the Behringer design is the fact that all the operational amplifiers are oriented the same way making the process MUCH easier. If you got one right, you got them all right.
I enjoyed that immensely. OPERATIONAL AMPLIFIER REPLACEMENT RESULTS
This was the LARGEST single contributor to the overall performance of the T1953. The amount of noise riding on the output signal was reduced by a factor great than 10... yes, you read that right, the amount of noise riding on the output signal was reduced by a factor great than 10. It cleaned up so nicely that the output wave actually looked like a square wave and not a triangle wave with hair on it, randomly pulsing. It was absolutely amazing how much of a difference those operational amplifiers made.
The output frequency response INVERTED. Previously, the T1953 was high-frequency impaired in that the square wave slanted toward the right. Now, it slanted left (low-frequency response). THIS IS GOOD!!! THIS IS GOOD.
I know, from reviewing Behringer's data sheet that the T1953's low-end 3dB point was 30Hz. The square wave actually showed that, correctly, now. This can be corrected by replacing the caps in the signal path - more on that later. This was the LARGEST SINGLE improvement that could be made to this box. The end result was stunning to say the least.
OTHER SIGNAL PATH UPGRADES
The next step, then was to figure out the last of the signal path upgrades that could be done WITHOUT costing an arm and a leg. I could have replaced all the Electrolytic caps, in the signal path with Black Gate caps. However, they would have cost me as much as the unit did, originally - silly idea, IMO. The goal is the most bang, for the buck - in other words, allowing significant performance improvements for each $$$ spent on the unit. To that end there are a few things that can still be done for the unit and improve performance. First thing you'll notice about this thing is that there are a LOT of electrolytic caps on the preamp PCB. There are 10 operational amplifiers, taking 20 of them for filtering, but there's something like 30 on the board.
First, how to determine the signal path electrolytics from the filter (power supply group). Pretty simple, really. The operational amps are powered with 15VDC. Typically, the rule of thumb is to use a cap with twice the voltage rating of the supply voltage. There are a bunch of 35V caps on the board - these are filter caps for the operational amplifiers.
There are a bunch of 10V caps on the board, too. These are SIGNAL PATH caps. Ahhh... that's how they do that. There are a couple of 60V caps, too. Those are for filtering the 48VDC phantom power.
At this point in time, I didn't have the time to spend to get me some non-polarized electrolytics for the signal path. Truth be told, that would provide some performance improvement, but not having another order ready for DigiKey, right now, made that "upgrade" a bit more $$$ than I wanted to spend - so I didn't do it.
The "problem" with the electrolytic capacitor is that it's really only a capacitor until about 5-6kHz, or so. Then it turns into an inductor and causes all sorts of problems. To rectify this, I got me a bunch more of the 1uF military grade surface mount caps and shunted all of the electrolytics in the signal path. This, again, significantly improved my high frequency performance.
WHAT IS THE UTC CIRCUIT?
I have no idea, really. Actually, interestingly enough, this circuit is a parallel tube (sort of like a parallel effects loop) that the signal is fed into. The signal is grossly distorted, the high-frequency harmonics filtered out and mixed into the signal path. Warmth? Not really, IMO, it works more like an exciter than something that adds warmth. It really adds a bit of "presence" to the signal that wasn't there before.
Interestingly enough, with the NJM4580's in the circuit the effect of the "warmth" wasn't hardly noticeable, even when I cranked the knob all the way to the right and the meter was pegged with all the warmth I was supposed to be adding. Now, however, the effect of the UTC circuit is VERY apparent. It adds a nice touch to the overall signal - especially voice, I've noticed.
THE FINAL RESULTS
That done, I took the unit out of the lab and took it home where I could actually test this thing in a studio environment with some relatively good mic's (RODE NTK, K2 and SHURE KSM32 and KSM44). I A/B'd the unit against my Groove Tubes VIPRE...
I have to say that, overall, the Groove Tubes and the T1953 are pretty hard to differentiate alone - you can tell a bit but not much. When they're in the mix, they both sound the same. The overall clarity, presence, dynamic range and depth of the T1953 had improved so much that it was nearly on par with a $3500 / channel mic pre. There's slightly more noise than the Groove Tubes - but hey, I spent a total of something like $50 in parts on this thing and $139. This is, now, a $200 product that comes close to sounding like a product greater than 10x it's price range.
WHAT UPGRADES MAKE SENSE ON THE BEHRINGER UNITS?
Due to the fact that Behringer uses surface mount components in everything there is some risk every time one performs an upgrade on these boxes.
I started out with high hopes of ordering every surface mount resistor and cap in some highly accurate, highly stable configuration to really improve the thing. I was going to order black gate caps, metalized film polypropelynes, metal film .5% resistors and the whole nine-yards.
The plain fact of the matter is that would have been very dangerous. Each of the surface mount devices has this leetle beety pad beneath it. That pad is the devices connection to every other device in its path. Those pads enjoy removing themselves politely from the PCB when heat is applied. Once they're removed you have to bridge between the component and the pad - making the task significantly more complex. Given the fact that Behringer actually glues every device to the PCB makes the chance of actually removing a pad significantly higher. In fact, on the T1953, I did remove a pad from one of the operational amplifiers. The glue was on the pad and when I removed the operational amplifier the pad came with it. The incremental value that you get for replacing components like this does not, IMO, offset the risk of replacement. The incremental performance upgrade doesn't offset the risk, either.
Upgrades of the signal path capacitors (electrolytics) make sense. Installing a same size, higher voltage (better performance) non-polarized cap makes HUGE sense. Performance improvement is there and it's all good.
Bypassing ALL electrolytics with smaller caps to improve high-frequency response - also a good thing. The surface mount caps really make this process nice. It improves high frequency response in the signal path AND improves RF filtering on the power supply.
Adding higher quality operational amplifiers to the signal path. This is the single greatest improvement in the Behringer product line that you can make. HUGE improvement gains can be had here. Adding FET input operational amplifiers, or other instrument grade operational amplifiers makes a HUGE difference in performance.
Adding RF shielding all the way around the chassis. Especially the cable packs. This makes a huge difference in the performance of low-level signals. With less EMI interfering w/ them, you'll be able to better hear the small transients and differences.
I have not pulled apart my DSP9024 ULTRADYNE, DSP8024 ULTRACURVE, ULTRAFEX PRO or the Virtualizer Pro. I'm betting that the performance of the digital gear can be upgraded, as well. Either with higher performance op amps feeding the A/D converters (that will help), higher performance op-amps feeding the outputs (that will help) better signal path components, better filtering on the power supply (stops A/D and D/A conversion glitches) and possibly even better A/D converters and D/A converters. However, that said, I haven't touched, or ever opened, one of their digital boxes. I will, eventually. I do know, however, that the ULTRAFEX PRO (all analog) will suffer the same performance improvements as what the mic pre's do when they're upgraded.
