ATtiny85 programmer for full design life cycle ?

I've just started looking in detail at the ATtiny85 and am applying to a specific project where I need to load a program into the ATtiny85, then use the serial monitor on the PC to look at the output mainly for debugging.
I'm not using a bootloader, simply a lightly hacked version of the standard example ArduinoISP sketch on a Uno clone which I've modified to be controlled with a single jumper to switch it from ISP mode for programming to a sort of "USB passthrough" mode so I can use the PC's serial monitor against the ATtiny85. This works fine, and is very simple and quick to switch over.

So far so good. Now, I have discovered that my application, a signal analysis program, does not work 100% reliably due to drift using the local oscillator at 16MHz so I have to dedicate 2 pins to a crystal and now have to consider using the reset pin as an analog input pin. This means I now have the additional requirement for a high voltage programmer so I can switch the function of the reset pin to a GPIO pin and back.

I can imagine doing the complete cycle: Fuse Resetting (High Voltage) -> program download to ATtiny85 -> Fuse setting -> "USB passthrough" for serial monitor -> testing, using a PC connected Uno or similar, and hopefully with a minimum of switching jumpers etc.

I have had a quick look around but have not found a complete solution although I'm sure other people have had the same idea. Before I start hacking out my own solution to incorporate a high voltage programmer (of which there are several examples around), I'd like to know if anyone has done something like this or have other useful information including design practices which allow for in-circuit programming of ATtiny85 without being impacted by, or impacting other installed components.

I don't have any particular comment on this situation. It's definitely awkward to have to reprogram the chip when you have to disable reset, particularly since you probably have to disconnect whatever is connected to the reset pin so the +12v doesn't damage it when doing HVSP...

Have you considered using an ATTiny84 (or one of the other bigger attiny's)? The Tiny84 (and slightly larger 861) are both available in through-hole packages, so they're just as easy to work with as the tiny85 - they just have more pins...

OK. Thanks. I'm lashing up a high voltage programmer now, but as a stand alone to see how it works. It'll be based on one of the examples, something similar to:

At the moment, it will be stand-alone, and later I'll work out a way of integrating it with the bits I already have with the aim of having a reasonably smooth development to deployment path for my ATtiny85 project(s).

Clear is that at least the reset pin will have to be jumpered so I can isolate it during the high voltage programming.

I'm aware of the 814 option, but I'm looking forward to the challenge of shoe-horning the application onto an ATINY85 even if I can no longer use the rather unique feature of the 85 that it has a 16MHz internal oscillator.

I'm doing my tests with a couple of ATtiny cores including yours. Karma ++ for your great contribution to this.

@Erni wrote a good summary of the debug options available...

http://www.ernstc.dk/arduino/tinycom.html

6v6gt:
So far so good. Now, I have discovered that my application, a signal analysis program, does not work 100% reliably due to drift using the local oscillator at 16MHz so I have to dedicate 2 pins to a crystal and now have to consider using the reset pin as an analog input pin. This means I now have the additional requirement for a high voltage programmer so I can switch the function of the reset pin to a GPIO pin and back.

How about using one of those 4 pin "TTL oscillators" and then pumping the clock signal into CLK1? That only consumes one pin instead of both pins if you use a plain crystal.

If you have space in memory you could use the bootloader somehow?

I'm resurrecting this topic after a long break. I was waiting a long time for a parcel of electronic stuff from China which I have now been using for this project.

Thanks for the suggestions. It is now clear to me that the Attiny85 is not appropriate for this activity. It is not the number of pins which I have solved. It is the quality of the ADC.
When I started detailed testing and noticed reliability problems, I also tried an Attiny84. Since that is does not have an internal 16MHz clock option, I added a crystal. The ATTiny84 did not have these reliability problems and indeed functioned perfectly for my application ( a AFSK demodulator). So I assumed it was to do with the external oscillator and it did not occur to me that the ADCs would be so different. That is until I looked further.

It appears that the data sheet for the ATtiny85 does say "65 - 260 μs Conversion Time" but also "Up to 15 kSPS at Maximum Resolution" which is a surprisingly high rate for such a slow conversion time. I require only 8 bit resolution at 9600 samples/second.
Interestingly, the datasheet for the ATtiny84 says "13μs Conversion Time" but also, surprisingly, "15 kSPS at Maximum Resolution".
In short, I've ruled the ATtiny85 out for this particular activity, however, I have learned a lot in the process.

Which leads me on to a by-product of my experiments with the ATtiny85 and trying to work around the limited number of pins leading to the use of the reset pin as a GPIO pin and the knock-on effects that has for program development. I have now built a programmer which combines the high voltage programmer necessary to restore the use of the reset pin, the ISP programmer for loading sketches, using AVRDUDE etc. and a USBTTL adapter for viewing the serial console etc. I've published it here: http://forum.arduino.cc/index.php?topic=508603.0
Maybe someone finds it useful.