John Warren Posted April 22, 2017 Share Posted April 22, 2017 Small signals handled in this area of the board. The op-amps (top left, bottom mid, bottom right SO-8 packages) are ST Micro MC33078 which was a surprise given the schematic shows NE5532s. Each package contains a couple of op-amps, so six total. The package on the lower left is an LM393 that's used to sense the present of the input signal. The entire board is RoHS compliant, the solder joints look like shit and many of the parts show evidence of soldering distress. There are a number of joints that are not SMD and appear to be done manually. Many of the mounted devices are nano-F range caps (0805 packages). . RoHS is a farce. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 22, 2017 Share Posted April 22, 2017 Here's the two caps that, should you wish, changing would be worth doing. These are coupling caps between the small signal gain and error amp (the flip side of the small signal section shown above). The factory uses the "RM" brand of polar caps seen in a lot of Chinese stuff. These are 10uF, 25V replace with Nichicon Muse. Quote Link to comment Share on other sites More sharing options...
minermark Posted April 22, 2017 Share Posted April 22, 2017 36 minutes ago, John Warren said: Amplifier board pulled from the amplifier That is a nice stack of Alum, im impressed. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 22, 2017 Share Posted April 22, 2017 4 hours ago, minermark said: That is a nice stack of Alum, im impressed. Agreed. The sinks are shrouded using a poly type of fiber reinforced "paper" which guides airflow across the fins. A couple of fans draw the air across the fins. The fans are quiet. The down side of the assembly is that it's near impossible to get at the output devices. At the factory, the output devices are attached to the heat sinks, the heat sinks then attached to the board (using machine screws that pass thru the bottom of the board into threaded holes in the sinks) and then the devices are soldered manually to the solder pads. So it really wasn't designed to be repaired a the device level. That said, the amplifier board is designed to be pulled and swapped for a new unit which is unfortunate because with a few design changes the devices could be relatively easy to replace individually. 1 Quote Link to comment Share on other sites More sharing options...
minermark Posted April 22, 2017 Share Posted April 22, 2017 2 minutes ago, John Warren said: Agreed. The sinks are shrouded using a poly type of fiber reinforced "paper" which guides airflow across the fins. A couple of fans draw the air across the fins. The fans are quiet. The down side of the assembly is that it's near impossible to get at the output devices. At the factory, the output devices are attached to the heat sinks, the heat sinks then attached to the board (using machine screws that pass thru the bottom of the board into threaded holes in the sinks) and then the devices are soldered manually to the solder pads. So it really wasn't designed to be repaired a the device level. That said, the amplifier board is designed to be pulled and swapped for a new unit which is unfortunate because with a few design changes the devices could be relatively easy to replace individually. Im no doubt like you when seeing adversity in design, the harder they make it to get in, the more challenge, but im going in. I do like hearing of good air flow, something more amps could use. Good luck with your device. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 22, 2017 Share Posted April 22, 2017 Replaced small signal coupling caps, Nichicon Muse (the green jacket ones 10uF, 25V). These are in the signal path between the gain and error amplifiers. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 22, 2017 Share Posted April 22, 2017 Chassis with amp board pulled. The Copper colored bosses are ground connections. The PC board holes at these locations are plated on both sides. The power supply board (upper right corner), the XLR/RCA input board (upper left corner) and transformer are all pretty easily pulled from the chassis. The headers and screws were covered in a red polystyrene glop that was there to keep them secure and (maybe) to indicate tampering. It's gone. Quote Link to comment Share on other sites More sharing options...
jwc Posted April 24, 2017 Author Share Posted April 24, 2017 I'm late to seeing your post. I'm sorry...I overlooked it. I'm grasping the graph a little. So you've shown the 10K Hz, power, distortion. But what are each of the individual lines of different color? IOW, why isn't there one line? So that is a board out of an XLI? Seriously? 1 Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 24, 2017 Share Posted April 24, 2017 1 hour ago, jwc said: I'm grasping the graph a little. So you've shown the 10K Hz, power, distortion. But what are each of the individual lines of different color? IOW, why isn't there one line? The plot shows %THD v. power sweep at 5 frequencies, the bottom is 100Hz, green is 1k, gold is 5k, light blue is 10k and dark blue 20kHz. Distortion is function of power and frequency. Suppliers, if they even publish a result plot 1kHz only. Distortion tends to increase considerably in the higher end of the spectrum (case in point). 1 hour ago, jwc said: So that is a board out of an XLI? Seriously? It's the amplifier board and includes both small and large signal stages. All on one board. So yes, I'm serious. Quote Link to comment Share on other sites More sharing options...
mikebse2a3 Posted April 25, 2017 Share Posted April 25, 2017 On 4/22/2017 at 10:23 AM, John Warren said: Some measurements. From bottom to top. plot is %THD at 100, 1000, 5000, 10k, 20k from about 1W to a little over 100W. The analyzer sampling is 192kHz. 20kHz test is hard test for amps to evidence numbers lower than say .5% at low power levels. The trends shown with power and frequency are typical. Thanks John...... interesting to notice how the amplifier's distortion is highest as power level decreases except at 100hz anyway. It would be interesting to see what happens with distortion in the milliwatts to 5watt range for those considering using these with high efficiency loudspeakers in an average home listening room. miketn Quote Link to comment Share on other sites More sharing options...
jwc Posted April 25, 2017 Author Share Posted April 25, 2017 I understand a little better now. Thanks. So what is your goal here besides "make better"? Mark1101 has bragged over and over with your work on his MC2100. I had considered many times getting a XLI800. I had no idea you knew the thing from bottom up and could maybe service it or "upgrade it" (not sure yet what you want to call it). Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 25, 2017 Share Posted April 25, 2017 8 hours ago, jwc said: So what is your goal here besides "make better"? It's a low cost class AB amp with rail switching and decent distortion numbers. It's straightforward to install an active, steep slope filter inside the unit and convert each channel into bass horn amplifier. Implementing a 4th order band-pass for, say an Klipschorn or LaScala, takes a PC board the size of a credit card. Quote Link to comment Share on other sites More sharing options...
jwc Posted April 25, 2017 Author Share Posted April 25, 2017 1 hour ago, John Warren said: It's a low cost class AB amp with rail switching and decent distortion numbers. It's straightforward to install an active, steep slope filter inside the unit and convert each channel into bass horn amplifier. Implementing a 4th order band-pass for, say an Klipschorn or LaScala, takes a PC board the size of a credit card. Wow, Please continue to share photos and explanation. Great stuff. Quote Link to comment Share on other sites More sharing options...
jwc Posted April 25, 2017 Author Share Posted April 25, 2017 You know, I'm realizing you had already unleashed the beginning of this info on this thread: Quote Link to comment Share on other sites More sharing options...
mark1101 Posted April 25, 2017 Share Posted April 25, 2017 13 hours ago, jwc said: I understand a little better now. Thanks. So what is your goal here besides "make better"? Mark1101 has bragged over and over with your work on his MC2100. I had considered many times getting a XLI800. I had no idea you knew the thing from bottom up and could maybe service it or "upgrade it" (not sure yet what you want to call it). John rebuilt both my MC-2100s. I have been incredibly happy with those and use them daily as 210 wpc monoblocks. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 26, 2017 Share Posted April 26, 2017 That's great to hear Mark! I own a couple myself, love them. The purist have issues with the basic design but with the right speakers the amp can make a lot of people smile. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 29, 2017 Share Posted April 29, 2017 On 4/25/2017 at 7:22 AM, jwc said: Wow, Please continue to share photos and explanation. Great stuff. The idea was to develop an active bandpass filter for any bass horn (say a Klipschorn) that's otherwise impractical to realize using passive components. The XLi is low cost, has excellent %THD below 1kHz, is class AB and lends itself to the mod. The actual filter is text book Sallen-Key, very easy to implement in the small signal section of the amplifier. In the schematic, the filter is outlined in red. I've placed it between the input op-amp (U2) and front panel pot comprised of R35, R44 (pot is at full CW). The schematic shows just the small signal processing of the XLi. U1 is the gain amp, U3 is the error amp. The response plots shown are small signal simulations that are possible using cascaded Sallen-Key filters. The PC board for the filter is installed vertically sharing a couple of the mounting screws used to mount the amplifier board to the chassis. By implementing a 4th order bandpass in the small signal domain the amp drives the Klipschorn to work within its horn loaded range. The horn loaded range is where it makes the best sound. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 30, 2017 Share Posted April 30, 2017 Here's how the rails switch in the XLi. Takes about half a second for the amp to turn on. The power supply takes about 150ms due to charging the filter caps. There's a power up relay that keeps the amp silent until its awake. In the plots, green is the high positive rail, red is high negative rail, light blue is the low negative rail and the pink is the low positive rail. The dark blue signal is 500Hz sine wave at the amp output, the input is 500Hz, 500mV balanced input from 500 Ohm source. On both the positive and negative rails, feedback from the output is sensed and used to pulse the gate of a power MOSFET which performs switching between low and high rails. As output voltage at the load increase, the duty cycle of the ON pulse increases. That's shown in the second plot. The switching enables high power without high idle dissipation. Quote Link to comment Share on other sites More sharing options...
John Warren Posted April 30, 2017 Share Posted April 30, 2017 Turning the wick up on the input signal from 500 to 800mV shows how duty cycle at switching increases. Note that the top rails are beginning to sag a bit. Quote Link to comment Share on other sites More sharing options...
John Warren Posted May 7, 2017 Share Posted May 7, 2017 The photo shows how the filter board was installed in the first units I modified. The bracket at the left is screwed into a chassis ground location and was used as the ground for the filter. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.