Parts Express sells the Yung SD series plate amps, which are nice light-weight class D amps that range from 100W to 500W. They use a switching power supply, they seem well made, and they all share the same preamp circuit board. However, their most interesting feature is that the preamp circuit board is more complex than most subwoofer plate amps, with a total of 3 quad op amps and one dual. In addition to the conventional 2-pole high pass filter (HPF) and adjustable low-pass filter, the amplifier uses a 2-pole multi-feedback band-pass (MFBP) filter to provide the bass boost. Furthermore, this circuitry is mounted perpendicular to the plate without a cover, so all components are easy to get to. In this project we will show how to modify the MFBP and HPF filters and how to model the response for use with different drivers and enclosures.
First, let’s take a look at the Yung preamp block diagram:
OK, the block diagram isn’t very easy to read, but if you copy and paste the image into Word you can actually read the text. Anyway, the real point is that there are a few more blocks in this this amp than what we usually see in a subwoofer plate amp. Usually a subwoofer plate amp provides boost by using a HPF with high Q, so that there is a peak in the response before the low frequency rolloff. However, with the Yung amps, the boost is implemented with a separate band-pass filter (MFBP) that can be adjusted separately. The schematics for these blocks are shown below.
These diagrams are the result of reverse-engineering the circuit board with an ohmmeter, and I think they are correct. However, I can’t provide a guarantee that I didn’t make a mistake. Also, the components shown are for the SD100-6 amplifier. Since the 3 other versions all have different boost frequencies, the components will be different, although the bare circuit boards appear to be the same. I don’t have any of the other amps to document the components, but if someone sends me an email or adds a comment to this page, we can start a table that shows the component values for each of the amps.
The HPF is a standard Sallen-Key configuration that is discussed in many online resources. The MFBP is also a common configuration that is well documented. A good resource that describes the theory and properties of both of these filters is this article: https://focus.ti.com/lit/ml/sloa088/sloa088.pdf.
In order to understand the overall speaker response, we are going to use a box modeling program that can also model the effect of the filters in the SD100-6 amp. The box modeling program is PSD-Lite, which has been around for many years. PSD-Lite is actually a full-featured loudspeaker design program that parses measurement files (“FRD”) for up to 3-way speakers, models box and baffle effects using driver TS parameters and baffle geometry, interactively allows designing a passive crossover, and models the effect of active filters in the woofer amp. The program was written for the .NET environment, using the free Visual Studio development tools. Although originally written for .NET version 3.5, it still compiles for .NET version 4.X, which allows updating the program without major impact to the existing code. The original code allowed modeling the HPF filter, and it was recently updated to model the MFBP filter as well. Another recent update is a crude but effective “optimizer” that determines the best set of resistor and capacitor values for these filters, based on maximizing the bass output over a specified range.
The next section will show how to use PSD-Lite to model the loudspeaker and enclosure and how to determine the “optimal” values for the filters. The PSD-Lite optimizer has a mode that only evaluates changes to the capacitors. The capacitors are through-hole components and the resistors are SMD on the SD100-6, so this will make modifications easier. The SMD resistors really aren’t that hard to replace if you have the right tools, but overall the through-hole capacitors are much easier to deal with.
Modeling Example Using PSD-Lite
<Need to make a better version of the video, using a larger background and a real script>
Modifying the Filters
First, you need to understand that modifying the amp will void the warranty. So consider how much you paid for the amp and whether doing these modifications is worth the loss of the warranty. Second, you should not do these mods unless you have the right tools. A temperature-controlled soldering iron is needed to ensure that you don’t damage the board, and you will need some fine braid or a solder sucker. And third, don’t let this be your first electronic modding experience. You should have some prior experience with replacing components on boards, as it is fairly easy to damage the printed circuit board if you haven’t done something like this before.
The SD100-6 amp is shown below. The preamp board is at the bottom of the picture, and those two pairs of blue capacitors nearest the two potentiometers are our target. As you can see, they are easy to get to and no disassembly is required.
The board is clearly labeled on the front–C4 and C5 are on the right side of the board in the picture below, and C13 and C14 are to the left of them, near the “Cutoff” control. Simply unsolder the existing components and replace them with film capacitors of the calculated value. The metallized polypropylene capacitors from EPCOS, Panasonic, Kemet or Vishay are all good choices, but make sure you get parts that fit. This means your selection will be limited to the 50V or 63V parts. I had some .33uF EPCOS parts that I used for C4 and C5, so I removed the existing parts and replaced them. But I didn’t have any of the .68uF parts that I needed for C13 and C14, so I just added some .22uF capacitors to the back of the board. The existing parts were .47uF, so the parallel combination came out close enough.
The picture below shows the .22uF capacitors added to the existing .47uF capacitors. I used a bit of hot-melt glue to keep them from vibrating, but that’s probably overkill.
And that’s it. It took many hours and days to update the PSD-Lite software to model the circuit, but that work is done. It will probably take about 20 minutes to set up the model, run the optimizer and re-run it until you are happy with the results. But the actual physical modification can be done in about 10 minutes.
And yes, it works. I used this for the “Act 1″ project posted at the Parts Express Project Showcase, for a small subwoofer. There isn’t a lot of room-filling low frequency output with the small 5-1/4” cones, but what’s there is deep and not “boomy”. Perfect for the intended application–a compact 2.1 system for computer audio (mostly near-field listening).